Journal Issue
Share
Article

Pakistan: Selected Issues and Statistical Appendix

Author(s):
International Monetary Fund
Published Date:
November 2005
Share
  • ShareShare
Show Summary Details

IV. Banking System and Stock Market Update15

A. Introduction

50. The 2004 Financial System Stability Assessment (FSSA) attested to the turnaround in Pakistan’s banking system (IMF Country Report No. 04/215). Seven years of restructuring, recapitalization, and privatization had by mid-2004 transformed the ownership structure, risk management, profitability, and reach of the system, allowing asset quality to improve and credit to expand into previously underserved segments of the economy. Nonetheless, the FSSA cautioned that rapid credit growth “could be problematic if sustained.”

51. This chapter updates some of the FSSA findings and tracks developments over the last 1½ years. Lending activity appears to have approached “boom” thresholds in the period since the FSSA, and the stock market underwent a contained boom-bust cycle in early 2005. Despite this, the banking system has continued to strengthen on an aggregate level, although trends at a few individual banks underscore the need for continued supervisory vigilance. Most financial soundness indicators (FSIs) remain on an improving trajectory. Nevertheless, bank-by-bank stress test results suggest that a few large banks are vulnerable to some shocks. The authorities have continued implementing their financial sector reform agenda and followed-up on FSSA recommendations. Development needs remain in the nonbank financial sector. On the stock market, several aspects still fall short of international good practice, as discussed in the FSSA, but ongoing reforms are moving in the right direction.

B. Financial Sector Overview

52. Pakistan remains a moderately intermediated economy with pervasive development needs that call for further financial deepening. Viewed through this prism, the rapid credit expansion of recent years is a welcome development. The spread of financial intermediation, in turn, has been facilitated by a wide-ranging, largely home-grown financial sector reform program, appropriately focused on the banking system. At end-2003, the latest year for which sector-wide data are available, total assets of the consolidated financial sector—defined here to include the Federal government’s direct deposit-taking operations through its National Savings Schemes (NSS)—stood at the equivalent of about 70 percent of GDP (Table IV.1). Money supply (cash, bank deposits, and NSS liabilities) amounted to 57 percent of GDP at mid-2005, compared with 87 percent for India (at end-March, including its Small Savings Schemes, which are essentially the same as Pakistan’s NSS; Figure IV.1).

Figure IV.1.Pakistan: Money Supply, 2005 1/

(In percent of GDP)

Source: National authorities, and Fund staff calculations.

1/ Latest available data.

Table IV.1.Pakistan: Financial Sector Assets, 2003
Billions of Pakistani rupeesPercent of GDPPercent share
Financial sector3,92270.9100.0
NSS 1/84415.321.5
Banking system 2/2,53845.964.7
Commercial banks2,43844.162.2
Public sector commercial banks95917.324.5
Local private banks1,21221.930.9
Foreign banks2674.86.8
Specialized banks1001.82.5
Development finance institutions791.42.0
Microfinance banks40.10.1
Nonbank finance companies1663.04.2
Leasing companies450.81.1
Investment banks250.50.6
Housing finance companies210.40.5
Mutual funds751.41.9
Modarabas 3/140.30.4
Pension funds1502.73.8
Insurance companies1262.33.2
General insurance companies300.50.8
Life insurance companies961.72.4
Sources: SBP; SECP; and Bank-Fund staff estimates.

Direct deposit-taking by the Federal government. Figures are for total liabilities, not assets.

Consolidates operations of foreign branches of domestic banks.

Islamic industrial finance companies.

Sources: SBP; SECP; and Bank-Fund staff estimates.

Direct deposit-taking by the Federal government. Figures are for total liabilities, not assets.

Consolidates operations of foreign branches of domestic banks.

Islamic industrial finance companies.

53. As in most developing countries, the financial sector is dominated by banks. Nonetheless, nonbank financial institutions play a significant supporting role. The banking system, which includes commercial banks and four small, state-owned specialized banks, accounts for approximately two-thirds of total financial sector assets. The banking system, as well as a small number of development finance institutions and microfinance banks, is regulated by the State Bank of Pakistan (SBP). The NSS account for another one-fifth of the financial sector and compete directly with banks for deposits. The remainder of the financial sector, including nonbank finance companies, pension funds, and insurance providers, is regulated by the Securities and Exchange Commission of Pakistan (SECP).

54. Eight years of reforms have transformed Pakistan’s banking system. Before reforms began in 1997, the system was dominated by chronically loss-making public sector commercial banks weighed down by substantial nonperforming loans (NPLs). Today, following a period of sustained restructuring, recapitalization, and privatization, the core of the system is made up of local private banks; assets are growing robustly; overall NPLs are falling even in nominal terms; and profitability is at record levels. The decline in the NPL ratio since 1997 has been helped by recovery drives, promulgation of a foreclosure law, restructuring of loans, issuance of write-off guidelines, and the takeover of some large NPLs by an asset management company (corporate and industrial restructuring corporation). Moreover, the flow of new NPLs has come down significantly. Led by consumer lending, credit is expanding into previously underserved segments of the economy without (thus far) undermining asset quality, and capital is growing faster than risk-weighted assets.

55. The ownership structure of the banking system has changed radically and competition has increased (Figures IV.2IV.3; Table IV.2). Following the privatization of United Bank Ltd. in 2002 and Habib Bank Ltd. in 2004, the market share (by assets) of public sector commercial banks had by end-2004 fallen to 21 percent, most of which was with the National Bank of Pakistan (the largest bank in the system). Between 1997 and 2004, the market share of local private banks increased from 26 percent to 65 percent, reflecting the two large privatizations, several acquisitions of foreign banks, and more rapid growth overall. At the same time, the market share of the five largest banks in the system dwindled from 62 percent to 56 percent. By assets and on several other measures, the market share of the next five banks increased marginally, and that of the remainder of banks by a more substantial amount.

Figure IV.2.Pakistan: Banking System Structure, 1997–2005

(Number of banks)

Source: Pakistani authorities.

1/ End-June.

Figure IV.3.Pakistan: Banking System Assets, 1997–2005

(In billions of Pakistani rupees)

Source: Pakistani authorites.

1/ End-June.

Table IV.2.Pakistan: Banking System Concentration,1997–2004 1/(Percentage shares)
19972004
Total assets100.0100.0
Banks 1–562.056.0
Banks 6–1015.017.0
Other banks23.027.0
Gross loans100.0100.0
Banks 1–560.453.8
Banks 6–1016.717.4
Other banks22.928.8
Customer deposits100.0100.0
Banks 1–567.059.9
Banks 6–1011.314.4
Other banks21.725.7
Net interest income100.0100.0
Banks 1–548.154.0
Banks 6–1028.118.2
Other banks23.827.8
Noninterest income100.0100.0
Banks 1–561.559.0
Banks 6–1010.018.8
Other banks28.522.2
Net income after tax100.0100.0
Banks 1–5130.358.5
Banks 6–10-10.821.8
Other banks-19.519.8
Source: SBP.

Commercial and specialized banks; consolidates operations of foreign branches of domestic banks. Banks ranked by total assets.

Source: SBP.

Commercial and specialized banks; consolidates operations of foreign branches of domestic banks. Banks ranked by total assets.

56. The system has been expanding rapidly since 2002. Reflecting improved capitalization and the macroeconomic turnaround, the average annual rate of growth of total assets doubled from 8 percent in 1998–2001 to 16 percent in 2002–04. With the asset growth of the public sector commercial banks and foreign banks fluctuating from year to year, the expansion has consistently been spearheaded by the local private banks, which have emerged as the most dynamic group in the system. Asset growth has closely been tracked by deposit growth in almost every year of the period, and the share of net loans (gross loans less allowances) in total assets has been rising since 2003.

C. A Credit Boom?

57. Credit growth in Pakistan has been rapid over the last two years but may not necessarily constitute a credit boom. Rapid credit growth can stem from financial deepening, cyclical upturns, or credit booms (Box IV.1). While there are some indications that a boom may be evolving, there are also fundamental reasons why credit should be growing rapidly without constituting too much risk. Strict adherence to prudential standards by banks and careful oversight by supervisors will be crucial to ensure that the momentum remains benign.

Box IV.1.Reasons for Rapid Credit Expansion (IMF, 2004)

  • Financial deepening. Financial intermediation moves in tandem with economic development and spurs economic growth. Thus, credit expansion can exceed economic growth, in particular as part of growth accelerations.

  • Cyclical upturns. Credit can also expand more rapidly than growth during an upturn because firms’ need for investment and working capital fluctuates with the cycle.

  • Credit booms. A credit expansion can become unsustainable if it is no longer based on future fundamentals. The financial accelerator mechanism can lead to such booms when shocks to asset prices are amplified by balance sheet effects. Balance sheet effects can also arise due to an increase in the relative price of nontradables, for example, in response to capital inflows.

58. International Monetary Fund (2004) discusses credit booms and assesses their risk for macroeconomic stability. A credit boom is defined as an episode where real credit growth exceeds the standard deviation of credit fluctuations around an HP-filtered trend by a factor of 1.75. In a cross-section of emerging markets, such booms are typically preceded by episodes of high real credit growth, in excess of 17 percent. The study concludes that credit booms pose significant risks because they are typically followed by sharp economic downturns and financial crises. A tightening of monetary policy is suggested to restrain credit growth and excessive demand, even if inflationary pressures are not building. However, framing an appropriate policy response is complicated by the fact that credit booms are not easily identified when they are happening.

59. Private sector credit is exhibiting some characteristics of a boom. Credit to the private sector grew by 19 percent in real terms in 2004 and has remained strong in the first half of 2005. Real private sector credit growth also touched the threshold for credit booms as defined above, and seems poised to exceed it in 2005 (Figure IV.4). Moreover, there are other “telltale” signs that typically accompany credit booms (IMF, 2004):

  • Investment has increased according to several indicators;

  • The current account shifted into deficit in 2004, after being in surplus since 2001;

  • The 12-month inflation rate reached 11 percent in April 2005 (before easing to 8.4 percent in August);

  • The stock market surged by 65 percent in the first 2½ months of 2005, taking price-earnings (P/E) ratios to as high as 20 (before a large correction in March);

  • Real estate prices reportedly increased by 70 percent in the six months to March 2005 (but have remained flat since); and

  • The share of net private sector credit in bank assets increased to over 50 percent in 2004 (and to more than 65 percent at some banks) and has continued to climb in 2005, from about 40 percent in 2002.

Figure IV.4.Pakistan: Private Sector Credit Growth, 1976–2004

(In percent) 1/

Citation: 2005, 408; 10.5089/9781451830675.002.A003

Sources: International Financial Statistics; and Fund staff calculations.

1/ Change in average annual credit to the private sector deflated by the period average CPI.

2/ HP-filtered trend plus the standard deviation of the cyclical fluctuation around the trend multiplied by 1.75.

60. Developments in Pakistan bear some semblance to typical crisis anatomies, but are more likely to reflect a growth takeoff (Box IV.2). The current economic boom was triggered by a political and policy change that restored confidence, followed by an easing of monetary policies and unexpectedly high private transfer that resulted in ample liquidity. Expectations became positive as economic growth accelerated, and the economy started to heat up—as witnessed, for example, by the acceleration in inflation, the asset price boom, and surging imports. But, put in historical perspective, the boom phase has not lasted very long, and the orderly correction of the Karachi Stock Exchange (KSE) in late March 2005 had virtually no impact on the economy and may have prevented a more severe bursting of a bubble later on.

Box IV.2.Crisis, Volatility, and Long-Run Growth

Kindleberger (2000) sketches the “anatomy of a typical crisis”. At the outset, there is an exogenous shock that results in a displacement and alters the outlook of economic agents. In response, agents try to take advantage of new opportunities. A boom commences, which is fed by an expansion of bank credit that enlarges the money supply. Eventually, some agents will engage in speculation, trying to benefit purely from rising prices. At some stage, insiders start to get out of the market, taking their profits. At the peak, a specific signal occurs that precipitates the crisis, for example, a bank failure. Agents try to liquidate their positions, which can become disorderly and turn into an outright panic or run.

Similar precrisis developments are identified in other studies. Allen and Gale (1999) identify financial liberalization and the ensuing credit expansion as the exogenous shock that leads to a displacement. Kaminski and Reinhart (1996, 1999) find that most crises are preceded by financial liberalization and credit expansion, followed by an average rise in equity prices of 40 percent per annum and a significant increase in real estate prices. An appreciating real exchange rate tends to precede an external crisis.

While bubbles and financial crises have severe ramifications, they may also be associated with strong growth performance over the long run. Ranciere, Tornell, and Westermann (2005) find that countries choosing a riskier growth path that involves credit booms and crises grow faster over the long run than countries that choose a less volatile growth path. As an example, they cite Thailand and India. The authors identify financial liberalization as triggering episodes of above-average growth that often end in crisis. However, the output gains during the high growth period more than offset the losses during the crisis.

61. While there are some warning signs, there are also fundamental developments that warrant strong credit growth. The recent credit expansion reflects financial deepening and a base effect. Reforms have yielded a more healthy banking system that is now well placed to lift financial intermediation above the low levels that had prevailed previously. In addition, with real credit growth having fallen to zero in 2002, the subsequent procyclical acceleration of credit as the economy rebounded from crisis was boosted by a base effect.

62. The broad-based nature of the credit expansion also suggests that vulnerabilities may be contained. The composition of net commercial credit by debtor shows credit to most sectors having grown at double digit rates since 2003, with exceptionally high credit growth witnessed in telecommunications, construction, and services (Table IV.3). In addition, consumer financing is growing by more than 100 percent, albeit from a very low base. The composition of net commercial credit by originating institution shows a similarly even distribution across the main bank groupings, with the notable exception of the specialized banks (which are mostly in workout mode). These developments suggest that the banking system may not have become more vulnerable to shocks in any specific sector and that risks are not likely to be concentrated in specific groups of banks.

Table IV.3.Pakistan: Banking System Net Commercial Credit, 2003–05 1/(Annual percentage change)
200320042005 2/
By type of bank:
Banking system28.934.732.0
Commercial banks36.343.735.7
Public sector commercial banks49.9-29.639.9
Local private banks45.386.035.6
Foreign banks-20.951.528.9
Specialized banks-3.3-20.4-0.8
By economic sector:
All sectors28.934.732.0
Agriculture and fishing16.17.118.8
Mining and quarrying160.4-34.0-30.4
Manufacturing42.127.727.8
Of which: textiles and garmentsN.a.N.a.30.6
Ship breaking and waste, scrap, junk, etc.-7.275.737.3
Construction20.248.271.7
Electricity, gas, water, and sanitation12.986.952.4
Commerce-0.867.240.6
Transport, storage, and communications13.5106.788.9
Of which: telecommunications112.6137.6195.4
Services-22.160.234.1
Other10.0105.460.4
Source: SBP.

Gross domestic loans, less specific allowances, to nonfinancial public sector enterprises and the nonfinancial private sector, excluding individuals.

End-June.

Source: SBP.

Gross domestic loans, less specific allowances, to nonfinancial public sector enterprises and the nonfinancial private sector, excluding individuals.

End-June.

D. Financial Soundness Indicators

63. Although credit growth has approached boom thresholds, there are no signs at present of compromised lending standards. Between mid-2004 and mid-2005, NPLs have fallen by 3 percent in nominal terms, even as total gross domestic loans have increased by 34 percent. The overall NPL ratio has declined by 4 percentage points, to 10 percent (Table IV.4). Importantly, the NPL ratio on credit to the corporate sector has declined by 6 percentage points, to 9 percent, underscoring the improvements in credit controls and risk management skills and practices at banks. Even the historically troubled area of agricultural credit has recorded a 5 percentage point improvement in asset quality, with its NPL ratio falling to 37 percent as of mid-2005. The one notable exception to the generally positive trajectory has been credit to small or medium enterprises (SMEs), where the NPL ratio increased by 4 percentage points in the 12 months to mid-2005, to 13 percent.

Table IV.4.Pakistan: Banking System Performing and Nonperforming Loans, 2004–05 1/
June 2004June 2005
Loans (% share)NPL ratio (%)Loans (% share)NPL ratio (%)
Gross domestic loans100.014.0100.010.1
Commercial credit80.116.377.012.5
Corporate sector54.914.652.38.9
SMEs17.29.317.413.1
Agricultural production8.042.27.337.0
Consumer finance7.61.011.40.9
Credit card advances0.83.61.11.4
Automobile loans2.50.63.70.7
Consumer durables loans0.17.10.16.2
Mortgage loans0.60.01.50.3
Personal loans and other3.60.65.11.0
Commodity operations6.71.27.81.2
Staff loans 2/2.91.32.21.4
Other 3/2.728.31.616.9
Source: SBP.

Domestic loans only.

Loans by banks to their own employees.

Includes loans to the Federal and provincial governments.

Source: SBP.

Domestic loans only.

Loans by banks to their own employees.

Includes loans to the Federal and provincial governments.

64. Asset quality has benefited from a sectoral shift in loan composition. Even as the NPL ratio on consumer finance has remained at or below 1 percent, the share of consumer finance in total gross loans has surged from 2 percent at end-2002 to 11 percent at mid-2005, a staggering average annual rate of growth of 142 percent. Almost half of consumer financing is in personal loans (mostly secured against salaries) and another one-third is in automobile loans (typically with 15–20 percent down payment, monthly installments, and a recovery rate of about 99 percent). The remainder of consumer finance is in mortgage loans (with down payments of as little as 15 percent, floating rates, and maturities of up to 10 years) and credit card advances. The NPL ratio on credit card advances had risen to 4 percent in mid-2004 but has fallen subsequently as customers began to find that delinquencies obstructed their access to other categories of credit. Banks report that consumer durables loans, as well as loans for two-wheeler vehicles, have grown less rapidly because of their higher ratios of administration cost to loan value.

65. Profitability has surged as a result of loan growth and efficiency gains, although applicable taxes remain high. In perhaps the single most remarkable indication of the restructuring of the Pakistani banking system, the ratio of noninterest expenses to total gross income (the “efficiency ratio”) has fallen from 85 percent in 1997 to 48 percent at mid-2005 (Figure IV.5). Before-tax returns on average assets (ROA) and average equity (ROE) reached a very healthy 2.3 percent and 35 percent, respectively, in the second quarter of 2005, while after-tax ROA and ROE reached 1.4 percent and 22 percent. The wide differential between pre- and post-tax profitability reflects a discriminatory corporate income tax rate (41 percent for banks vs. 35 percent for other corporations—although the differential has been reduced over the last three years and will be eliminated in 2006) and the non-tax deductibility of loan-loss provisions (a departure from international best practice, as pointed out by the FSSA).16

Figure IV.5.Pakistan: Banking System Efficiency and Profitability, 1997–2005

(In percent)

Source: Pakistani authorities.

1/ End-June

66. Strong profitability and capital injections have buffered the system against unexpected losses. In 2004, for example, Allied Bank Ltd. received a cash injection of PRs 14.2 billion ($238 million) from its shareholders. The “raw” capitalization ratio (capital and surplus to total assets) for all banks increased from 3.5 percent in 1997 to 6.6 percent at mid-2005 (Figure IV.6). More impressive still was the improvement in the capital adequacy ratio (CAR, total regulatory capital to risk-weighted assets), which increased from 4.5 percent to 10.9 percent during the same period, with only modest reliance on subordinated debt, unrealized gains on investments, and other “supplementary” capital. Capital adequacy guidelines have been progressively tightened: Basel I norms were instituted in 1997; risk-weighting misclassifications by banks were corrected in 2003; and country risk guidelines as well as capital charges for market risk were introduced in 2004.17

Figure IV.6.Pakistan: Banking System Capitalization, 1997–2005

(In percent)

Source: Pakistani authorities.

1/ End-June

Box IV.3.Pakistan: Absolute Capital Requirements and Provisioning Norms

The SBP has used absolute capital requirements as a tool to influence system structure. The minimum paid-up capital requirement for banks has been increased from PRs 1 billion to PRs 1.5 billion with effect from end-2004 and PRs 2 billion ($34 million) by end-2005. Of the 39 banks in operation at mid-2005, nine had yet to achieve the PRs 1.5 billion figure. Despite this, the SBP has announced its intention to further raise the capital floor, to the rupee equivalent of $100 million by around 2009, in order to weed out those small banks that it views as providing few banking services while imposing a significant regulatory burden. It should be noted, however, that the absolute capital floor in the Euro zone, often viewed as a benchmark, is only €5 million; that small banks can be a key source of innovation; and that mandatory equity injections risk leaving banks with a choice between depressed ROE or rapid (often excessively rapid) credit growth to lift ROA.

Specific provisioning requirements remain lenient. Minimum criteria for NPL recognition are adequate, with loans required to be classified as “other assets especially mentioned” (“OAEM”) when interest or principle falls 90 days past due. “OAEM” do not, however, carry specific provisioning requirements. “Substandard,” “doubtful,” and “loss” loans do carry such requirements, but a medium- or long-term commercial credit, for instance, need not be classified as “loss” until 3 years past due. International good practice would suggest that loans be classified as “substandard”, “doubtful”, and “loss” at 90 days, 180 days, and 1 year past due, respectively. In the current environment of falling NPLs and record profitability, banks are well positioned to absorb tighter specific provisioning requirements, which in turn would better prepare them for possible future problems. Secured and unsecured consumer financing carry cautious general provisioning requirements of 1.5 percent and 5 percent, respectively.

67. Loan-loss allowances have also grown strongly. Prudential guidelines laying out minimum standards for the classification and provisioning of commercial credit, SME credit, and consumer finance were issued in 2003. The guidelines include objective criteria governing the forced sale value of collateral, with a three-year phase-in period. Overall, the NPL coverage ratio (total allowances to gross NPLs) increased from 47 percent in 1997 to 74 percent at mid-2005, while the ratio of net NPLs (gross NPLs less allowances) to capital and surplus declined from 184 percent to 24 percent (Figure IV.7). As of end-2004, no less than 87 percent of the commercial banks’ gross NPLs consisted of fully provided loans classified as “loss” (banks show a reluctance to write off “loss” loans and thereby risk extinguishing legal claims, and are allowed to keep such loans on their balance sheets for up to three years).

Figure IV.7.Pakistan: Banking System Asset Quality, 1997–2005

(In percent)

Source: Pakistani authorities.

1/ End-June

68. As loans have expanded, liquidity has tightened. The ratio of net loans to total assets increased from 46 percent in 1997 to 52 percent at mid-2005 (Figure IV.8). As a natural corollary of the portfolio reallocation in favor of loans, the ratio of liquid assets (cash and balances with the SBP and other banks, call money and repurchase lending, and investments in government securities) to total assets fell from 40 percent to 36 percent during the period.18 The tightening of liquidity thus far is unlikely to be problematic, especially given the relatively deep interbank “call money” market in Pakistan.

Figure IV.8.Pakistan: Banking System Liquidity, 1997–2005

(In percent)

Source: Pakistani authorities.

1/ End-June

69. In sum, FSI aggregates for the banking system point to impressive and ongoing improvements in capital adequacy, asset quality, and earnings (Table IV.5). Indeed, the number of banks with CARs below the minimum required level of 8 percent had by end-2004 fallen to just one, from five only three year earlier. The specialized banks as a group are an exception to the rule, however, with the Industrial Development Bank of Pakistan remaining in a negative net worth position pending restructuring and privatization.

E. Stress Test Methodology and Results

70. Stress tests complement the FSI-based findings.19 Stress tests were conducted to assess whether improving FSI aggregates mask any erosion of capital adequacy at individual large banks such that they would be less prepared to absorb (fairly severe) potential shocks. The exercise was conducted based on the FSSA methodology, with bank-by-bank data for September 2003 (the FSSA data set) and end-2004 (Table IV.6). Bank-specific data were confined to the 12 largest commercial banks (by assets), with the specialized banks excluded from the exercise. Six shocks were simulated (Box IV.4).

Table IV.5.Pakistan: Selected Banking System FSIs for Middle East and South Asia, 2000–04 1/(In percent)
20002001200220032004 2/
CAR:
Bahrain23.426.426.526.730.3
Bangladesh6.76.77.58.48.8
Egypt10.210.29.911.011.1
India11.111.411.912.913.4
Iran6.66.94.57.2
Jordan19.417.416.715.917.8
Kuwait22.222.019.718.4
Lebanon16.918.019.422.3
Nepal4.0-7.3-11.7
Oman16.515.617.117.617.6
Pakistan9.78.88.88.510.9
Saudi Arabia21.020.318.719.418.0
Sri Lanka8.27.89.410.49.1
United Arab Emirates19.419.618.618.216.3
GCC average20.520.820.120.120.6
South Asia average9.27.75.24.511.0
Gross NPL ratio:
Bahrain11.211.611.212.68.9
Bangladesh28.122.117.6
Egypt13.615.616.920.224.2
India12.811.410.48.86.6
Iran4.34.45.45.75.2
Jordan18.419.321.019.713.6
Kuwait19.210.37.86.1
Lebanon7.810.012.412.812.2
Nepal29.330.428.7
Oman7.510.611.315.513.5
Pakistan23.523.421.817.010.6
Saudi Arabia10.410.19.25.43.1
Sri Lanka13.615.315.313.710.0
United Arab Emirates12.715.715.314.312.5
GCC average12.211.711.010.89.5
South Asia average18.221.422.719.211.6
ROA after tax:
Bahrain1.81.81.51.92.3
Bangladesh0.70.50.50.7
Egypt0.90.80.70.50.5
India0.70.50.81.01.2
Iran0.90.6
Jordan0.30.70.50.71.0
Kuwait2.02.01.82.0
Lebanon0.70.50.60.7
Nepal-0.7-3.7-4.3-1.4
Oman1.30.21.50.31.9
Pakistan-0.2-0.50.11.11.4
Saudi Arabia2.02.22.32.32.5
Sri Lanka0.80.71.11.41.2
United Arab Emirates1.81.91.91.92.1
GCC average1.81.61.81.72.2
South Asia average-0.1-0.8-0.70.31.1
Sources: Pakistani authorities; and Fund staff estimates.

Figures for Pakistan are for commercial and specialized banks, consolidating operations of foreign branches of domestic banks.

Latest available. Figures for Pakistan are for the second quarter of 2005.

Sources: Pakistani authorities; and Fund staff estimates.

Figures for Pakistan are for commercial and specialized banks, consolidating operations of foreign branches of domestic banks.

Latest available. Figures for Pakistan are for the second quarter of 2005.

Table IV.6.Pakistan: Stress Test Pre-Shock Summary Statistics for the 12 Largest Commercial Banks, 2003–04 1/(In percent)
September 2003December 2004
Weighted averageSimple averageStandard deviationWeighted averageSimple averageStandard deviation
Risk-weighted assets to total assets36.141.711.555.258.813.0
CAR11.911.39.310.711.13.2
Gross loans to regulatory capital964.3717.3638.5940.8964.4205.8
Gross NPL ratio17.611.911.79.97.36.9
Composition of gross NPLs100.0100.0100.0100.0
OAEM2.96.08.74.15.28.1
Substandard4.56.14.34.14.77.1
Doubtful7.711.214.15.011.318.0
Loss84.976.615.186.978.722.6
Allowances to gross NPLs 2/62.459.616.974.3114.2135.6
OAEM 3/0.26.319.08.66.319.0
Substandard 3/14.514.07.516.014.07.5
Doubtful 3/27.327.319.523.727.319.5
Loss63.958.630.078.673.614.8
Collateral to gross NPLs 4/35.542.220.722.026.715.1
OAEM0.00.00.00.0
Substandard27.938.138.820.432.435.1
Doubtful45.451.138.452.946.736.9
Loss36.243.525.521.426.414.8
Remaining time-to-repricing gap to regulatory capital 5/118.2109.219.1302.4320.591.8
t ≤ 3 months-347.7-248.9471.5-63.0-28.7195.9
3 months < t ≤ 1 year426.2275.0395.9153.8135.9269.7
t > 1 year39.783.2449.3211.7213.3172.2
Net open foreign currency position to regulatory capital-36.9-16.939.925.20.861.7
Gross foreign currency loans to regulatory capital 6/170.2121.592.9155.5110.378.9
Total equity exposure to regulatory capital34.514.534.821.819.615.4
Equity investments20.410.818.514.512.811.5
Carry-over-transaction financing2.02.43.83.23.46.6
Equity collateral 7/12.01.217.64.13.42.9
Real estate collateral to regulatory capital 8/48.24.770.316.513.611.6
Sources: SBP; and Fund staff estimates.

Banks ranked by assets; consolidates operations of foreign branches of domestic banks.

Includes general allowances; September 2003 ratio of general allowances to total allowances assumed to be identical to the (actual) ratio for December 2004.

Specific allowance coverage ratios for September 2003 assumed to be identical to the (actual) ratios for December 2004.

Collateral values imputed by assuming that specific allowance coverage ratios conform with regulatory requirements.

Remaining time-to-maturity gap for September 2003.

Actual data for September 2003. December 2004 figures assume gross foreign currency loans constitute 50 percent of gross foreign currency assets.

Equity assumed to constitute 20 percent of total collateral.

Real estate assumed to constitute 80 percent of total collateral.

Sources: SBP; and Fund staff estimates.

Banks ranked by assets; consolidates operations of foreign branches of domestic banks.

Includes general allowances; September 2003 ratio of general allowances to total allowances assumed to be identical to the (actual) ratio for December 2004.

Specific allowance coverage ratios for September 2003 assumed to be identical to the (actual) ratios for December 2004.

Collateral values imputed by assuming that specific allowance coverage ratios conform with regulatory requirements.

Remaining time-to-maturity gap for September 2003.

Actual data for September 2003. December 2004 figures assume gross foreign currency loans constitute 50 percent of gross foreign currency assets.

Equity assumed to constitute 20 percent of total collateral.

Real estate assumed to constitute 80 percent of total collateral.

Box IV.4.Pakistan: Details of Stress Test Shocks

  • Regulatory risk. The first shock sought to gauge the preparedness of banks for a potential tightening of specific provisioning norms. It was assumed that NPLs classified in the unprovisioned “OAEM” category would be reclassified as “substandard” (with an obligatory minimum coverage ratio of 20 percent); those in the “substandard” category would become “doubtful” (50 percent); and those in the “doubtful” category would become “loss” (100 percent).

  • Credit risk. The second shock assumed a 35 percent nominal increase in gross NPLs, with the incremental NPLs classified as “doubtful” (and provisioned at 50 percent).

  • Interest rate risk. The third shock assumed a steepening of the yield curve, with the effective interest rate for assets and liabilities due to reprice within 3 months increasing by 100 basis points; that for assets and liabilities due to reprice between 3 months and 1 year increasing by 300 basis points; and that for assets and liabilities due to reprice beyond 1 year increasing by 500 basis points.

  • Exchange rate risk. The fourth shock sought to gauge the preparedness of banks for the direct and indirect effects of a potential rupee depreciation. In addition to the direct valuation effects of a 10 percent weakening of the Pakistani rupee vis-à-vis the U.S. dollar on the net open foreign currency position of each bank, it was also assumed that 20 percent of unhedged gross foreign currency loans would be reclassified from performing to “doubtful” (provisioned at 50 percent).

  • Equity price risk. The fifth shock assumed a 30 percent decline in the value of equity investments and equity collateral, with 30 percent of equity-related financing reclassified from performing to “doubtful” (provisioned at 50 percent). It was further assumed that equity constitutes 20 percent of imputed collateral.1

  • Real estate risk. The sixth and final shock assumed a 50 percent decline in the value of real estate collateral, assumed to constitute 80 percent of imputed collateral.

1Minimum specific allowance coverage ratios are set as a percentage of gross classified loans less collateral (booked at forced sale value). Accordingly, with specific allowance data provided by each NPL subcategory for end-2004, it was possible to impute collateral values by assuming that actual allowances conformed with required amounts.

71. Test results reveal a mixed picture and underscore the need for continued supervisory vigilance (Table IV.7). Resilience to a regulatory tightening may have fallen marginally, but vulnerability to a (severe) credit shock is broadly unchanged—although, as noted below, the results are not fully comparable. Vulnerability to an interest rate shock has increased somewhat, as has vulnerability to an exchange rate shock or an equity price shock (such as that witnessed in March 2005). Resilience to a property price crash, conversely, has improved as loan-loss allowances have increased and reliance on collateral, including real estate collateral, has wound down.

Table IV.7.Pakistan: Stress Test Results for the 12 Largest Commercial Banks, 2003–04

(No. of undercapitalized banks; percent deposit share in parentheses) 1/

2003 2/2004
One-category reclassification of NPLs1 (5.9)2 (8.3)
35 percent increase in NPLs 3/4 (37.2)4 (35.1)
Yield curve steepening 4/3 (10.9)3 (25.0)
10 percent rupee depreciation 5/1 (5.9)3(11.9)
30 percent stock market decline1 (5.9)2 (8.3)
50 percent real estate market decline3 (27.5)3 (25.0)
Sources: SBP; and Fund staff estimates.

Undercapitalization defined as CAR < 8 percent. Results for 2003–04 not strictly comparable.

End-September.

Incremental NPLs provisioned at 50 percent.

Rate increase by 100 b.p. for t ≤ 3 mo.; 300 b.p. for 3 mo. < t ≤ 1 yr.; and 500 b.p. for t > 1 yr.

With 20 percent of foreign currency loans becoming NPLs, provisioned at 50 percent.

Sources: SBP; and Fund staff estimates.

Undercapitalization defined as CAR < 8 percent. Results for 2003–04 not strictly comparable.

End-September.

Incremental NPLs provisioned at 50 percent.

Rate increase by 100 b.p. for t ≤ 3 mo.; 300 b.p. for 3 mo. < t ≤ 1 yr.; and 500 b.p. for t > 1 yr.

With 20 percent of foreign currency loans becoming NPLs, provisioned at 50 percent.

72. Strict comparability between test results for 2003 and 2004 is hampered by a number of factors. First and foremost, CARs for September 2003 are overstated because banks had until then been misclassifying loans collateralized by commercial real estate (with a required risk weight of 100 percent) as if they were collateralized by residential real estate (with a risk weight of 50 percent); the resulting understatement of risk-weighted assets was corrected during the fourth quarter of 2003, but the stress test results for September 2003 were unrealistically favorable look better than they would otherwise be. Second, the test for a regulatory tightening used an imputed distribution of specific loan-loss allowances for September 2003 (it was assumed that specific allowance coverage ratios by NPL subcategory were the same as at end-2004). Third, the test for interest rate risk used maturity gap data for September 2003 and (more accurate) repricing gap data for 2004. Finally, the test for exchange rate risk used foreign currency loan data for 2003 and foreign currency asset data for 2004 (it was assumed that unhedged foreign currency loans constituted 50 percent of gross foreign currency loans at the first test date, and 25 percent of gross foreign currency assets at the second test date).

F. Equity Market Developments and Reforms

73. The stock markets had a rollercoaster ride in the first quarter of 2005 (Figure IV.9). Much of the buoyancy since 2002 can be attributed to the strong economy and the privatization program in addition to ample liquidity stemming from easy monetary conditions and inflows from abroad. However, the 65 percent price increase during the first 2½ months of 2005 (with P/E ratios reaching 20 at the peak) and the subsequent 30 percent correction in the latter half of March has been blamed on speculation and market manipulation, prompting an official inquiry. The inquiry found that a small number of key players had first fuelled prices by injecting liquidity into the “ready” (or spot) market and engaging in illegal “wash trades” (simultaneous purchase and sale of the same share to create an impression of active trading) and then depressed prices by withdrawing liquidity, after locking in high selling prices in the futures market.20 Following the March correction the market has stabilized (with P/E ratios now at 12) and several regulatory actions have been taken. Banks’ exposure to the stock market was and remains limited so that their balance sheets were not significantly affected by the correction.

Figure IV.9.Pakistan: Karachi Stock Exchange 100 Index, 2002:1–2005:8

(2001:12=100)

Sources: Pakistani authorities; and Fund staff calculations.

74. The equity market is modest in size, highly speculative, and “mutualized” in structure (Box IV.5). KSE market capitalization is about 37 percent of GDP (Figure IV.10). The benchmark index is the capital-weighted KSE-100, although trading is mostly confined to six public sector companies in oil, telecommunications, and banking. KSE trading concentration (the proportion of traded value accounted for by the ten largest companies) was 67 percent in 2004 (compared, for instance, with 29 percent for the Nasdaq); turnover velocity (traded value as a proportion of market capitalization) averaged 304 percent in 2004 (250 percent for the Nasdaq) and reached a very high 630 percent in the first quarter of 2005; the settlement ratio (settlements as a proportion of traded value) averaged only 7 percent in the first five months of 2005; and the free float is estimated at about 20 percent. Settlement in the ready market is on a t+3 rolling cycle.

Figure IV.10.Pakistan: Stock Market Capitalization, 2005 1/

(In percent of GDP)

Sources: Pakistani authorities; Bloomberg; and Fund staff calculations.

1/ Latest available data.

Box IV.5.Pakistan: Demutualization of the Stock Exchanges

Pakistan’s three stock exchanges are owned by market participants. The KSE (the largest and oldest), the Lahore Stock Exchange, and the Islamabad Stock Exchange are all “mutualized”, i.e., owned by their members, most of whom are brokers, creating a conflict of interest in their role as self-regulatory organizations. The KSE, for instance, has 200 members, 155 of whom are active in the market, and a member-controlled Board of Directors. The SECP is the apex regulator. Custody services are provided by the Central Depository Company (CDC), and clearing and settlement services by the National Clearing Company.

Demutualization of the exchanges is likely to prove contentious. In neighboring India, which inherited a similar securities trading architecture as Pakistan, the traditional hegemony of the stock broking establishment over the colonial-vintage stock exchanges was broken only through the launch and subsequent rapid growth of a new National Stock Exchange. The SECP has consulted with the Securities and Exchange Board of India on the issue. The Lahore Stock Exchange and the Islamabad Stock Exchange appear amenable to some form of merger, but the KSE remains reticent.

75. Financing in the ready market has until recently been dominated by the uniquely South Asian “badla” or carry-over-transaction (COT) system. Badla was a form of post-trade financing that resembled a repurchase transaction with equity collateral. Shares were traded each morning, and financing arranged in a one-hour sitting each afternoon. Badla rates were capped at 18 percent per annum at the KSE (increased to 24 percent during the recent boom) but were uncapped at the Lahore Stock Exchange (where they rose to over 100 percent during the boom). The exchanges facilitated the badla financing by serving as a platform, but were not involved as counterparties. Badla providers were mostly brokers and to a lesser extent banks, with the five largest and ten largest financiers providing about 50 percent and 75 percent of total badla financing, respectively. Total badla financing peaked at PRs 40 billion ($674 million) in mid-February 2005 before falling to PRs 27 billion ($455 million) at end-March (Figure IV.11).

Figure IV.11.Pakistan: Total COT Financing, 2002:7–2005:6 1/

(In billions of Pakistan rupees)

Source: Pakistani authorities.

1/ Latest available data.

76. Badla carried added risk for the clearing houses because it was post-trade financing. To cover this risk, the exchanges imposed additional margin requirements of 5–11 percent on badla-financed trades, over and above their standard margin requirements of 5–20 percent. In the event of default by a broker, he would first forfeit his collateral, then his assets would be liquidated, and any residual losses to the clearing house would be met from the Clearing House Protection Fund (up to PRs 50 million per default) and the Investor Protection Fund (up to PRs 10 million per member); as of end-March 2005 the two funds had assets of PRs 684 million ($12 million) and PRs 386 million ($7 million), respectively.

77. The chief problem with badla was that it facilitated very high levels of investor leverage—up to 100 percent—at very high interest rates and was provided without adequate due diligence. Instead, inordinate emphasis was placed on the perceived “guarantee” provided by the exchanges via the Clearing House and Investor Protection Funds. A second problem was that the concentration of lenders, most of whom were also market participants on a proprietary basis, allowed manipulation of market liquidity, as allegedly was the case in early 2005. Finally, abuses also reportedly took place, with some brokers raising financing for themselves by posting shares deposited with them by their investors; this was possible to the extent that such shares were pooled together and held in “group accounts” at the CDC, with insufficient clarity on beneficial interests. To close this loophole, group accounts have been eliminated, with brokers required (since April 2005) to maintain separate, client-specific subaccounts at the CDC. Experience with the CFS will be reviewed in early 2006.

78. The SECP recognizes the intrinsic problems with the COT system and has been unsuccessfully trying to phase out badla since 2002. Badla financing was limited to 30 listed companies with effect from mid-December 2003, with the number of eligible companies to be gradually reduced to zero by June 2005. The intention was to replace badla by margin financing from banks, with the margin requirement set at 30 percent. The phase-out was suspended in April 2005 following protests from brokers citing illiquidity concerns.

79. In August 2005, a compromise was agreed under which badla was replaced with immediate effect by a new Continuous Funding System (CFS). The CFS has been described by some observers as a modified badla, and characterized by some within the SECP as a one-year transitional arrangement ahead of a more fully developed futures market. The main difference between badla and the CFS is that under the latter the financing session will remain open all day, eliminating the post-trade-financing aspect of badla and hence reducing the risk to the clearing houses. There will be a new CFS account at the CDC, and the total volume of CFS financing is capped at PRs 25 billion.

80. The futures market remains embryonic and its further development will be central to efficient price discovery. Futures are confined currently to 30-day single-scrip contracts that resemble ready-market transactions with t+30 settlement, as opposed to modern futures contracts in lots of, say, 1,000 shares with clear terms and conditions. Futures settlement is based on deliverables only, because cash contracts are considered unIslamic. The exchanges, as well as some market participants, have emphasized the need for deeper derivatives markets, arguing that in India it was futures, not margin financing, that replaced badla. Accordingly, plans are being formulated to introduce 60-day and 90-day (single-scrip) futures, as well as a new 30-share Sensitive Index at the KSE, to be followed by index-linked futures and other derivative products.

81. In sum, reforms are systematically addressing those aspects of Pakistan’s stock markets that still fall short of international good practice. With several brokerage houses reportedly having taken losses in March 2005, it seems likely that the orderly market correction has served as a timely “wake-up call”, adding to the impetus for all stakeholders to work together to modernize market structure and trading practices.

References

Prepared by Ashok Vir Bhatia (MFD) with contributions from Axel Schimmelpfennig (MCD).

SBP proposals to make provisions tax deductible have met with longstanding opposition from the Central Board of Revenue.

The introduction of country risk guidelines and capital charges for market risk in late 2004 filled two important lacunae identified by the FSSA assessment of Pakistan’s compliance with the Basel Core Principles of Effective Banking Supervision.

Government securities in banks’ held-to-maturity accounts are eligible for repurchase with the SBP and are therefore counted as liquid assets.

The SBP has integrated the FSSA stress testing methodology into its supervisory toolkit. Univariate and multivariate stress tests are conducted quarterly, with summary findings published in the SBP’s “Quarterly Performance Review of the Banking System”.

Other Resources Citing This Publication