6.1 Overview

While policy focus on strengthening the regulatory framework continued apace during the year, it was the phenomenal rise in the deposit base, on the back of improvements in the country's external sector, that was the major highlight of the banking sector's performance during FY02.

Overall deposits rose by 14.0 percent, the strongest growth recorded since FY97, which was all the more impressive given the decline in foreign currency deposits (the 13.7 percent fall in Rupee value of the latter was masked by a startling 19.7 percent rise in rupee deposits). As explained in Chapter 5, the tremendous increase in the deposit base was generated by the unprecedented improvement in Pakistan's current account October 2001 onwards, and the corresponding increase in SBP foreign currency purchases.'

By contrast, growth in net credit was relatively subdued during FY02, but not as weak as suggested by the monetary survey data, as becomes visible only after adjusting for factors that artificially depressed the reported net credit figures (see Section 6.3.2). As a result of the available liquidity and expectations of declining interest rates, banks bid aggressively in auctions for government papers (see Section 5.10 on Money Market for details). Another very interesting development, also stemming from the external sector improvement, was the considerable interest in foreign currency loans; by end-FY02, the outstanding foreign currency loans reached to a stunning Rs 20.4 billion (US$ 339.5 million), accounting for a quarter of the net credit growth during the year.2 Finally, FY02 also saw visible indications of improvements in banks' Non-performing Loans (NPLs); the decline in NPLs (against domestic operations of banks), that began during the second half of FY01, not only continued but also gained momentum. The gap between the NPLs and defaulted loans has started to narrow, indicating at the better quality of fresh loans.

6.2 Policy Environment

The thrust of the regulatory framework during FY02 was much the same as in the previous fiscal year, with policies focusing on increasing the ability of the banking system to weather financial shocks, strengthening their capacity to extend credit at lower cost, reducing the direct role of the government, and fostering good governance. The following sub-sections discuss key policy developments during the year.

The banking industry in Pakistan remains dominated by 3 nationalized and 2 partially-privatized banks, despite the emergence of private sector competitors following the banking sector liberalization initiated in the early 1990s. These bigger banks, for the last several years, have been confronted with structural drawbacks - especially overstaffing, unprofitable branches, poor customer services and inept credit discipline. To minimize the organizational weaknesses of nationalized banks and to improve the financial soundness of the overall system, privatization of these units have become

Table 6.1 : Foreign Currency Loans (Stocks)

Source: Exchange and Debt Management Department, SBP

inevitable. To facilitate this process, a vigilant restructuring of these institutions has remained on the agenda over the last five years. The rationalization of staff and branch network, establishment of CIRC to acquire their NPLs, reduction in tax rates, and promulgation of privatization ordinance are some of the measures taken to ensure the speedy process.

In September 2001, a 6.4 percent government stake in MCB has been off-loaded and only 10.2 percent more shares remain to be divested. The process for privatization of UBL through sale of a 51 percent stake was initiated in June 2002 and is now in its final phase. In addition. Privatization Commission (PC) has invited Expressions of Interest (EOI) from the investors interested in acquiring a minimum 26 percent stake in Habib Bank Limited. Also, the government offloaded 10 percent shares of National Bank of Pakistan through initial public offering in the local stock exchanges and has decided to sell another 10 percent in coming months.

In line with international practices, a risk-based capital adequacy system is already in place since 1997.3 However, during FY01 SBP doubled the minimum paid-up capital (net of losses) requirement for scheduled banks to Rs 1.0 billion.4 Banks were allowed to meet the requirements in two phases, with a minimum capital requirement ofRs 750 million to be met by December 2001 and the final target to be achieved by end-December 2002.

The measure aimed to encourage economies of scale and to further strengthen the competitive ability of banks. Banks failing to meet this minimum requirement will be converted into non-scheduled banks with corresponding restrictions on their banking activities. The first phase concluded in December 2001 (for progress, see Section 6.4.1). In addition, banks have been allowed to issue Term Finance Certificates (as subordinated debt) to raise capital; this will help them improve capital adequacy.

Since there is risk aversion by banks towards extending credit to small and medium enterprises (SMEs), credit constraints had limited the expansion of these businesses. Realizing the large potential for growth and employment opportunities offered by the small-scale sector, government encouraged the creation of specialized institutions to extend credit to this market segment. As a result, a Small and Medium Enterprise (SME) Bank was established in public sector w.e.f January 1, 2002.

Similarly, in order to provide increased access to small borrowers, particularly in rural areas, two other micro-credit institutions Khushali Bank and First Micro Finance Bank, started operations. In recognition of their anti-poverty focus and the nature of their operational requirements the SBP relaxed the licensing and regulatory environment for these institutions.

Finally, FY02 also saw the emergence ofMeezan Bank, Pakistan's first commercial bank operating purely on Shariah-compliant transactions (for details on development in Islamic banking see Chapter 10 on Islamization of Financial System in Pakistan).

As mentioned earlier in Chapter 5, SBP took an easy monetary stance throughout the year and the interest rates declined accordingly. The EFS rate that is now being adjusted on a monthly basis, was brought down to 8 percent from 14 percent at the start of FY02. In this context, banks and exporters have been advised to avail the benefits of rate cuts and re-price their outstanding amounts. Also relaxations have been given, by extending maturities, to certain industries such as the carpet and rugindustry. New items have also been made eligible for refinance facility, including brown rice, information technology software as well as consultancy services such as medical, pharmaceuticals, engineering, etc.

A three-pronged approach has been adopted to address the issue ofNPLs. First, the Corporate and Industrial Restructuring Corporation (CIRC) was established in September 2000 to shift the ownership ofNPLs from banking institutions. The activities of CIRC gained momentum in FY02. Second, the Committee on Revival of Sick Units (CRSU) was established. The CRSU has restructured and revived 130 units, which had stuck-up loans with banks. Third, the cases of willful defaulters have been referred to NAB after following due process of law. Banks have been asked to increase emphasis on adequate provisioning for their NPL portfolio. Also, in order to remove the legal difficulties and time delays faced by banks in recoveries against defaulted loans, the Financial Institutions (Recovery of Finance) Ordinance, 2001 has been promulgated. It ensures expeditious recovery of stuck up loans by foreclosure and the sale of mortgaged property, with or without the intervention of courts.

A number of measures have been introduced in FY02 in this regard:

• To ensure transparency in banks' financial statements, they are required to report data on new formats in accordance with international standards.

• Banks have become liable to take prior clearance of SBP for appointment of CEOs.

• For promoting good corporate governance and to encourage an effective role of CEO/BOD, certain guidelines under "Fit and Proper Test" have been formulated for their appointment. These guidelines include conforming to the parameters like honesty, reputation, experience, track record, management and financial integrity.

• In order to ensure sound banking practices and existence of proper checks and balances in each institution, SBP has been issuing clear lines of responsibilities and proper guidelines to the board of directors. They have been made responsible to review and update policies in areas like internal audit, compliance, risk management, credit disbursement, management information system, etc. so as to enhance effective governance in the financial institutions.

• Also, the panel of auditors, that play an important part in promoting good governance in the banking sector, is being revised from time to time so as to improve the quality of audit services.

Following the mushroom growth in large denomination Rupee Travelers' Cheques (TCs) that were feared to be fostering undocumented transactions in the economy, the SBP capped the denominations of these instruments at Rs 10,000. Moreover, details on issuance of TCs are required to be reported to the SBP.

Taxation has remained a major issue for the banking industry over the years. The tax rates applicable to Pakistani banks have been significantly higher than for other corporates. Additionally, many financial transactions face withholding taxes, the payments on which often cumulate to levels above the total tax liability of banks. While these do appear on banks' balance sheets as tax assets, the cash loss is obviously a drag on profitability (for details see Special Section 6.1). Also, banks received no tax exemptions for accrued (but unrealized) interest on non-performing assets. These issues are often cited by banks when defending their inability to further reduce lending rates.

The corporate tax rate that had been cut from 58 percent to 50 percent in July 2001, was again lowered to 47 percent in the FY03 budget. The rates will be further reduced over the next four years

to bring it at par with the rates applicable to other corporates. Also, the withholding tax on T-bills is being reduced; and Rs 22 billion was paid in June 2002 as a part-payment against outstanding advance taxes.

Finally, provisioning costs and interest income taken to the suspense account have been exempted from taxable income. All these relaxations in tax liabilities of banks will improve the shape of their balance sheets, enabling the banks to narrow the wedge between deposits and lending rates.

Banks saw a remarkable expansion in deposits in FY02, recording a rise of 14.0 percent on an year-on-year basis. To put this in perspective, this growth is the strongest for any year since FY97 (see Figure 6.1), and in terms of Rupee deposits, it was one of the best since the introduction of Resident Foreign Currency Deposits (RFCDs) in 1991. As with many other macroeconomic variables in FY02, the catalyst for this change appears to be the post-September 11 improvements in Pakistan's external account. Specifically, a burgeoning current account surplus permitted the central bank to greatly increase its foreign currency market purchases without an adverse impact on the exchange rate (for details see Chapter 9 on Balance of Payments). This US$ 3.9 billion purchased by the SBP during FY02, represented a massive injection of fresh Rupee liquidity into the domestic economy. Not surprisingly, the growth rate for Rupee deposits doubled in FY02 to an incredible 19.7 percent against the 8.5 percent growth observed in FYO 1. It must be noted that following the introduction of the FE-25 deposits, the rise in foreign currency deposits largely did not result in an increase in banks' Rupee liquidity, hindering credit expansion in the domestic economy. By contrast, the Rupee appreciation catalyzed by the current account surplus not only led to a US$ 195.9 million (5.8 percent) fall in foreign currency deposits but also depressed the Rupee value of the remaining deposits. As a result, in Rupee terms, these deposits fell 13.7 percent during the year (see Figure 6.2).

As visible in Figure 6.3 the FY02 deposit growth was initially in line with the FY01 trends, and the September-October 2001 decline, which is visible for both Rupee and Foreign Exchange (FX) deposits, seems attributable directly to the September 11 shocks.5 The strong surge in deposits that followed in the next two months seems to be based on the sharp improvement in the current account balance (particularly remittances) and the corresponding increase in SBP FX purchases. In fact, this increase in deposits would have been even stronger, had there not been the negative impact of the fall in foreign currency deposits. By mid-FY02, the substantial appreciation of the Rupee, and expectations of further strengthening, had eroded the earlier "devaluation expectations"6 of foreign currency holders sparking the liquidation of their holdings.

The deceleration visible in H2-FY02 is explainable primarily by the sharp increase in mobilization under National Saving Schemes (NSS), the seasonal decline in credit offtake,7 and an accelerating decline in foreign currency deposits (FCDs). In fact, it seems that with FCDs no longer a favored avenue, savers focused instead on the other familiar option, the NSS instruments. Not surprisingly, the surge in NSS investments is a mirror of the fall in FCDs, helping the government to raise Rs 76.1 billion from this source against Rs 40.7 billion mobilized in FY01 (see Figure 6.4 & 6.5).

The structure of growth in deposits also makes an interesting comparison with FY01. Firstly, the deposit gains ofRs 173.5 billion in FY02 were spread across the year, in contrast to FYO 1 in the last month of the year.8 The increase in FE-25 deposits during H2FY02 in the face of the decline in overall FCAs (see Figure 6.5) can partially be explained by the announced merger of the FE-31 deposits with the FE-25 category.9 The rising level of lending in foreign currency may also have a role behind rising deposits (FE-25), through the multiplier effect. Also, an unprecedented increase in kerb market inflows temporally drove the kerb rate below the interbank rate, allowed the arbitrage. Finally, the currency composition of the incremental deposits under the FE-25 reveals another possible explanation: anecdotal evidence suggests an increased focus on Euro and Sterling denominated accounts owing to their appreciating values. Within the banking sector, the NCBs emerge as unambiguous leaders in deposit generation accounting for 54.7 percent of the overall FY02 growth (see Table 6.2). However, private banks maintained the highest growth rate.10 The striking success of the NCBs in raising deposits during FY02 is quite encouraging in view of the relatively poor performance in recent years. An important contribution to this could have been the recent efforts to attract remittances through official channels,11 aided by their substantial capacity to absorb external inflows.12 Foreign banks, on the other hand, saw deposits decline by Rs 16.3 billion.13 There are a number of reasons behind this, but the principal factor seems their relatively higher exposure to foreign currency deposits. Firstly, these comprised 30 percent of their total deposits, and thereby these institutions were particularly vulnerable to the impact of the appreciating Rupee, i.e. the exchange rate impact as well as the loss of deposits. Also, in recent months most foreign banks were unable to generate incremental FE25 deposits as they were already in breach of the SBP limits on deposits mix.14 In other words their FX deposits primarily comprised of older US$ deposits, which saw the biggest decline within FCAs.

6.3.2 Total Credit by Scheduled Banks In contrast to the phenomenal surge in deposits and declining lending rates, the broad credit numbers for FY02 look surprisingly weak. Specifically, the net credit15 extended by banks saw an increase of only Rs 41.8 billion compared to Rs 66.9 billion recorded during FY01 (see Figure 6.6). This becomes even more surprising given that net credit extended in FYO 1 was in the face of a severe shortage of liquidity and increasing interest rates. One explanation forwarded for this apparent anomaly is the fall in business confidence in the aftermath of 9/11, the uncertainty due to conflicts in Afghanistan, and the tension with India. However, though these negative shocks certainly put a dampener on economic activities, the 3.6 percent increase in real GDP (against a target of 4 Table 6.2 : Deposits by Scheduled Banks (Flows)

Source : banking Supervision Department

Incorporating a reasonable 4.4 percent growth in manufacturing, and sharp jump in the quantum of exports, suggests that the decline in the "headline" net credit figure may be misleading.

Some possible explanations were discussed in the SBP Third Quarterly Report tor FY02 including a fall in the prices of key inputs, higher CBR tax refunds,16 loan write-offs, etc.17

In fact detailed analysis demonstrates that factors such as: (1) higher deletions/write offs during FY02 compared to FY01, (2) shifting ofRs 22.1 billion loans from the schedule banks to government liability, (3) ban on credit extension to PSEs and Autonomous bodies in March 2002, and (4) larger disbursement of fresh loans during the year, taken together, provide a more comprehensive picture than the reported (lower) net credit demand figures. Thus, in order to more correctly assess the net credit picture, the subsequent analysis is based on the credit adjustment for the first two factors as well as the impact of mergers, for both FY01 and FY02. The adjusted figures show that FY02 net credit off-take was practically unchanged from that in FY01 (see Table 6.3).

The availability of funds to the non-government sector looks even better; the Rs 17.2 billion in incremental FY02 tax refunds would have obviously depressed credit demand from banks accordingly. It is also worth noting that only a part of the rupee injection made through net SBP kerb forex purchases was captured by the banking system, as evident in the sharp rise in currency in circulation. A portion of the latter balances could have also helped finance business activity, reducing the need for credit from banks. It is important to note that cumulative net credit (adjusted) started picking up late in FY02 compared to FY01 (see Figure 6.7). This subdued growth till the end of Ql -FY02 may be explained by: (1) higher retirement due to larger credit extension in FY01 over FY00, (2) decline in credit under EFS, and (3) the immediate impact of uncertain conditions after 9/11. It is generally observed that higher credit extension in a fiscal year results in higher retirement in the proceeding one. Accordingly, the nominal growth in net credit during Q 1 - FY02 despite significantly higher gross disbursement probably also reflected this phenomenon. Finally, Q 1 -FY02 posted a net contraction of Rs 12.1 billion in EFS credit against an expansion ofRs 1.2 billion in the corresponding period a year earlier

Table 6.3: Adjusted - Net Credit by Scheduled Banks (Flows)

Source: Banking Supervision Department, SBP

Thereafter, net credit off-take reflected normal seasonal demand trends. The incredible increase ofRs 58.2 billion in net credit during December 2001 is of particular interest though as: (1) all banks participated in the rise, and (2) weighted average lending rates were at a low during this period.

The subsequent decline in net credit appears seasonal, but it too incorporates interesting developments. First, a prosaic explanation for the unusually sharp fall in net credit in April-May 2002 could be that the bar on incremental bank financing for PSEs and autonomous bodies in March 2002 held down the net credit extension in the period.20

More interesting however is the spectacular growth in foreign currency denominated loans in the latter halfofFY02 (see Figure 6.8).21 These loans began rising gradually, in line with increasing expectations of a stable (or appreciating) Rupee, as the forex loans were extremely attractive in the absence of the exchange rate risk, with nominal interest rates ranging from 2-5 percent compared to much higher Rupee lending rates.22

Not surprisingly, by end-June 2002, the outstanding forex loans totaled Rs 20.3 billion in Rupee terms. To put this into perspective, this is approximately 25 percent of the total adjusted net credit growth during the year.

A look at the performance of the banks by segments also provides interesting insights given their differing credit cycles. It appears that NCBs and privatized banks mainly finance the activities that follow a seasonal pattern; their net credit starts picking up in September, peaks in December and then follows a declining trend. On the other hand, private and foreign banks primarily finance trade related activities (their comparatively higher shares in EFS also support this view) that continue through the year.

During the first five months ofHl-FY02, nationalized banks posted an exception to the declining trend in overall net credit. On the other hand, while net credit by privatized banks started increasing only from the beginning of Q2-FY02 (see Figure 6.9), their cumulative figure for HI -FY02 remained negative. It is important to note that in this period privatized banks saw low deposit growth. Both, nationalized and privatized, banks saw a sharp upsurge in net credit during December 2001, though this was much steeper for the former. The

Table 6.4 : Net Credit by Scheduled Banks (Flows)

Source: Banking Supervision Department

large jump for the nationalized banks may be linked to a 146 basis point decline in their weighted average lending rates and their larger share in incremental deposits mobilized by the banking sector. During H2-FY02, though the net credit of nationalized and privatized banks followed seasonal declines till May 2002, but this was much sharper than the previous year, probably due to the restrictions on financing ofPSEs and autonomous bodies. Nevertheless, NCBs managed to end the year with slightly higher increase in net credit during FY02 over previous year. Privatized banks on the other hand, recorded a negative growth (see Table 6.4).

The net credit growth of private banks in FY02 depicted a behavior at variance from that of the overall banking industry (see Figure 6.9).23 The decline in EFS credit appears to be the prime reason for their negative credit extension during the first five month of FY02, and the subsequent improvement appears to reflect their greater participation in foreign currency lending.

Finally, foreign banks largely failed to expand their credit during FY02, and particularly the sharp fall in H2FY02 appears to reflect a depleting deposit base.

The declining trend in NPLs that started after Q3-FY01, was sustained during the current year, but this becomes visible only after adjusting for a one-off jump in February 2002 (see Figure 6.10).24 Defaulted loans (a sub-set of NPLs) followed the same trend-line until HI-FY02, but thereafter saw a slight upward drift.

Although the increase in defaulted loans is not a welcome sign, indicating that at least some NPLs had worsened, the continuing overall decline in NPLs is unambiguously positive, suggesting that banks incremental loan portfolio has fewer problems.25

The two prime factors were responsible for this improvement:

• Fresh loans were being made in prudently. In FY02, Rs 73.2 billion were new additions against Rs 88.2 billion of previous year.

• Cash recoveries were higher in FY02. Specifically, Rs 33.7 billion were recovered during the year against Rs 30.4 billion in FY01. Particularly heartening to note is the sharp decline in NPLs of nationalized banks, which was partially driven by the transfer of bad loans to the CIRC. As the share of these banks is very high in total bad loans of scheduled banks, an improvement in this segment was badly needed. The slower decline in the NPLs for the banking sector as a whole reflects the deteriorating portfolios of specialized and few small banks in private sector.

The easing of monetary policy during FY02 did trickle down to the lending and deposit rates of banks, but largely during H2FY02. Also, the weighted average lending rates saw a larger decline compared to that of the weighted average deposits rate. As a result, the banking spread narrowed by 111 basis points to 7.9 percent (see Figure 6.11).26

This decline in spread became possible because of: (1) a cut in the tax rate to 50 percent with effect from July 2001, from 58 percent a year earlier;27 (2) a fall in non-performing loans;

(3) a relatively liquid interbank market and low reliance on costly borrowing (compared to the pervious two years) due to the phenomenal growth in deposits; and 4) a downward revision ofEFS rates by 600 basis points.28

As depicted in Figure 6.12, the weighted average lending rate declined primarily after November 2001, even though the discount rate, and the benchmark T-bill yield, had dropped sharply much earlier.29

A factor in this could have been liquidity problems in the interbank market, as suggested  by high overnight call rates until November 2001. This argument is supported by the fact that bank deposit growth accelerated only after October 2001.

Table 6.5: Financial Indicators of Scheduled Banks

percent

Source: Banking Supervision Department, SBP

The following section will analyze the financial health of the banking sector during CY01. The data and findings of this section differ from those in the earlier sections mainly due to three reasons. First, the data in this section is based on annual audited balance sheets of banks at their end financial year (December) and covers the global operations of banks, while data in last section pertained to domestic operations only. Second, the analysis here is based on calendar year while the objective of earlier section was to cover developments during FY02 following the overall structure of the report. Third, the segmentation of the banks in this section is different from Section 6.3. The Bank-wise list in their respective grouping of this section is given in the Statistical Annexure 6.12.

On a December-to-December basis, banking sector continued to grow during CY01. Net assets (net of provision) of the overall banking industry registered an increase of 7.3 percent over last year to Rs 1939.6 billion, which was funded entirely through a surge in the deposit base that saved banks from resorting to expensive borrowings.30 While this enabled most banks to improve profitability, many others, especially public sector banks, remained burdened by hefty provisioning requirements. In fact, the aggregate losses of the two public sector banks outweighed the aggregate profitability of the other banks. As a result, the financial indicators of profitability and capital adequacy of the overall banking industry deteriorated. However, appropriate provisioning, deletions & write-offs, and effective recovery drives, improved the asset quality of banks (see Table 6.5).

The deadline for banks to meet the first phase of enhanced capital requirements was end- CYO 1 (banks had been directed to raise their respective paid-up capital - net of losses - to Rs 750 million).31 Developments on this front show a mixed picture. Twelve out of total 43 banks have not achieved the target of which half have been exempted, and for the other half, deadline has been extended. Nevertheless, the paid-up capital of the banking sector witnessed an increase ofRs 6.2 billion (7.5 percent) during CY01 over the previous year (see Figure 6.13). Most banks issued both, right and bonus shares, to meet the requirement. It may be noted that aggregate paid-up capital also increased in FY01, but this was primarily because of a capital injections ofRs 8.0 billion by SBP in one of the public sector banks.

Despite a modest increase in aggregate paid-up capital, the total equity of the banking industry declined by Rs 9.2 billion (11.2 percent) over last year (see Figure 6.13). This fall primarily stems from losses suffered by two public sector banks due to substantial provisioning against the large historic stock of un-provisioned non-performing assets. While the increase in the aggregate provisioning is a positive for the risk profile of the banking sector, it also leads to deterioration in the capital adequacy indicators, as visible in the CY01 data (see Table 6.5).

The SBP requires banks operating in Pakistan to maintain the risk-based capital at not less than 8 percent of risk-weighted assets, in accordance with BASEL committee recommendations and in line with international practice. Despite a CY01 deterioration, on aggregate, the capital to risk- weighted assets ratio (CRWA) remained above this benchmark (see Table 6.5). However, if public sector banks are excluded, the figures see an improvement, since the entire decline this year is attributable to them. In fact, the ratio for the public banks has already gone below the 8 percent level in CYOO, and it slipped further in CY01 (see Table 6.6).32 This was largely due to one of the NCBs that had a substantial shortfall in equity in CY01, which was subsequently covered through capital injections by SBP.

Nevertheless, still there is a need to improve the capital base of banks. A NPLs to capital ratio of over a hundred percent is still a serious threat to the equity of banks. Efforts are going on in this direction: banks are passing through the second phase of capital enhancement and several steps have been taken to deal with chronic NPLs.

The large stock of non-performing loans of banks (primarily of public sector banks) is the biggest drag on their performance. On the one hand, banks are not earning any income on these assets, and on the other hand, provisioning against such loans is reducing profitability. Thus, dealing with bad loans has remained an priority in the on- going reforms process. To address this problem, several measures were taken, which included recovery drives, different incentive schemes and the establishment of CIRC. Though the stock of non-performing loans of the banking system is still very high, their pace has slowed down. It is essential to mention that NPLs in this section pertain to banks global operations, hence may differ from that given in Section 6.3.3 against domestic operations only. Furthermore, figures and analysis here are based on annual audited balance sheets of banks at end December, while the earlier section extended the discussion till end FY02.

Comparing the end December NPLs in 2001 and 2000 reveals an apparent deterioration of assets portfolio. Specifically, the non-performing loans of the overall banking system surged by 5.2 percent over the last year. However, the figures need adjustments on two counts: (1) the merger ofNDFC (a non-bank financial institution) with the NBP33 shifted bad loans of former to the banking sector; (2) the transfer of operations of UK branches of some of the banks to their newly established subsidiaries. The merger ofNDFC with the NBP resulted in amassing another Rs 13.6 billion to the stock of NPLs of the banking system, whereas transfer of UK branches operations contributed around Rs 6.5 billion. Adjusted figures exhibited a comparatively subdued

percent

Source: Banking Supervision Department, SBP

Increase of 2.2 percent against much higher growths of 4.0 percent and 26.1 percent in CY99 and CYOO respectively.

It is important to see the growth in NPLs in relation with that of advances. In CYO 1, the gross NPLs to advances ratio of the banking industry remained almost unchanged at the CYOO level (see Figure 6.14). Net NPLs to net advances ratio improved in CY01. A group-wise analysis reveals that public sector and foreign banks were better off whereas domestic private banks saw little change (see Table 6.7). Seen at a bank-wise level, the median value of this ratio came down in 2001 compared to the previous year. More specifically, 50 percent of total banks had net NPLs to net advance ratio either at 8.4 percent or below in 2001 against 9.5 percent a year earlier.

With the objective of strengthening the financial institutions, banks have been encouraged by SBP over last couple of years to make appropriate provisioning. The outturn of these efforts is not only increased absolute figures of provisioning but also its ratio to NPLs (see Table 6.5).

Analyzing the ageing of non-performing loans, it is apparent that major part fell in the loss category (especially for banks in public sector). As the future recovery of such long-outstanding loans appears unlikely, it may not be desirable for the banks to carry forward these on their balance sheets particularly when they hold provisions there against. Write-offs will not only improve the balance sheets of the banks but also lead them to focus more on fresh lending. SBP has recently issued guidelines to this effect, which will help the banks to clean up their balance sheets by writing off such un-collectible loans.

Higher provisioning dragged down the profitability of the whole banking sector in CYO 1. During the year around 12.0 percent of total income was eaten up by provisions as compared to 8.3 percent last year. Overall, banks' aggregate pre-tax profit recorded a sharp fall to Rs 1.1 billion against Rs 4.5 billion of previous year. After tax profit portrayed a more dismal picture, as banking industry suffered huge losses ofRs9.8 billion during CY01 against losing Rs 2.8 billion in previous year. However, this deterioration was not broad based. The hefty losses sustained by two public sector banks, primarily because of massive provisioning against their NPLs, turned the whole sector in red. If adjustment is made for the two banks the profitability of the banking sector shows a remarkable improvement during CY01. This could be seen with Rs 7.1 billion after tax profit during current year compared to Rs 2.8 billion losses of FY00.

Domestic private banks made a significant break-through by registering apre and post tax profits of Rs 5.0 billion and Rs 2.0 billion, respectively (see Table 6.8). This was in sharp contrast to previous

Percent

Source: Banking Supervision Department, SBP

Table 6.8: Earnings and Profitability

Source: Banking Supervision Department, SBP

year, when this group recorded a pre-tax loss ofRs 0.6 billion that turnout to Rs 3.5 billion after deducting the tax amount. Foreign Banks faced problems during second half of CYO 1, but still managed to improve after tax profits to Rs 2.4 billion against Rs 1.4 billion earned in CYOO. Though most of public sector banks (six out of nine) substantially increased their profits in CY01, exceptionally large losses suffered by two banks in the group dragged the profitability of whole sector into losses. As discussed earlier losses were mainly because of higher provisioning by these banks.

Higher profit earned by most of the banks during CY01 was primarily driven by higher net interest income. For the overall banking system, this rose to Rs 63.4 billion in CY01 from 46.5 billion of last year. On the expenditure side, re-profiling of the funds helped the banks in maintain the interest expenses for CY01 almost at the preceding year's level, despite a substantial increase in deposits. This was made possible because of increased share of deposits in current and saving categories (relatively cheap) during CY01, consequently the share of term-deposits has gone down. In addition, given the exceptional deposit growth, especially in local currency, banks resorted less to borrowings in CY01. On the other side increase in interest income, primarily on account of growing volume of earning assets also played important role in raising net interest income and the net interest margin (see Table 6.5)

Calendar year 2001 ended with ample liquidity in the interbank market. This was in sharp contrast to the previous year when banks faced one of the most severe liquidity shortages in the banking history of Pakistan and overnight rates skyrocketed. Although the banks were indeed struggling to replenish their deposit base (especially in local currency), the CYOO turnout was primarily triggered by SBP efforts to meet tough NDA targets for end-December 2000. In CY01, both of the above problems were absent; the remarkable surge in deposits towards the end of CY01 significantly improved liquidity positions of banks, and developments on the external account in the aftermath of 9/11 made the end period NDA target much easier (for detail see

Chapter 5 on Money and Credit). As a result, all the liquidity indicators showed improvement

during CYO 1, and the declining trend in the liquid assets to total assets ratio, visible since 1998, was reversed (see Figure 6.15).

The improvement in liquidity can also be seen through the fall in the loan to deposit ratio. Significantly higher growth in deposits compared to advances was the prime factor behind decline in loan to deposit ratio during current year. Given the excess liquidity that was available during the closing months of the CY01, interbank borrowing was expected to go down; it did, but only by 2.0 billion. This, while surprising, is attributed to foreign banks' appetite for funds, on the back of their depleting deposit base during the second half of CYO 1 (see Section 6.3.1). A good part of the increase in foreign banks asset came from borrowings that went up by Rs 17.0 billion or 30.0 percent over the year. On the other hand, other groups were improving as evident from rising liquid assets to borrowing ratio.

Notwithstanding the above positive developments, the maturity profile of assets and liabilities showed considerable mismatches in shorter period of time (up to three-months). More specifically, the liabilities maturing within three-months period significantly exceeded the maturing assets in this period by Rs 132.1 billion. This, however, is understandable, as with the expectations of further fall in interest rates on government securities banks naturally preferred to book their assets into longer-term maturities (for detail see section on Money Market).

Equity risk is not significant for banks simply because of the very small share it occupies in the investment portfolio as well as in total assets of the banks. However, banks are facing interest rate risk. The gap profile reveals that the rate sensitive assets and liabilities of the banking system, maturing within a three months time frame have a negative gap of Rs 277.6 billion or 14.3 percent of the total assets (see Figure 6.16). Having negative gap in the nearest bucket is natural in the declining interest rate scenario, as it would have a salutary impact on the earnings of the banking system. The standalone as well as cumulative gaps, as percentage of total assets, however, are within the internationally accepted limits of ±10 percent in all the buckets, except up to three months time frame wherein the exposure is slightly above the acceptable limit.

The group-wise position shows that the interest rate exposure of public sector banks up to three months at negative 23.5 percent and up to one year at negative 14.1 percent exceeds the normal limit. The exposure of other two groups remained within the limit of ±10 percent in all buckets.

The prime objective of the financial sector reforms, that started in late 1980s, was improving the efficiency of the financial sector.34 Given the significantly higher share of banks in overall financial system of Pakistan,35 efficiency of financial sector can hardly be achieved without its prior accomplishing for banks. The gap between lending and deposits rates (banking spread) is generally being used as a proxy of intermediation cost and hence its lower value is usually interpreted as improvement in banking efficiency. Contrary to the expected results of structural reforms, the actual data showed that the banking spread has almost doubled during the 1990s.36

Specifically, the gap between weighted average lending and deposit rates has widened from 4.8 percent in June 91 to 8.3 percent at the end of FY01 (see Figure 1).37 This widening of spread is often cited to conclude that the reforms were a failure. However, this is not a correct interpretation. The increase in banking spread should be interpreted with caution, as in the pre-reform period, interest rates were controlled from both ends, with floors on deposit rates and ceilings on lending rates. Their widening after the reforms partially indicated the change from a repressed to a liberalized interest rate regime. Moreover, it is important to note that the banking spread might not be a good indicator for gauging the efficiency of the banking system, as it does not cover all the interest earning and interest paying activities of banks. The interest rate spread (a broader definition) may be used as a better alternative for this purpose.38 As shown in Figure 2, the interest rate spread also increased during the 1990s (from 4.0 percent in 1991 to 4.9 percent in 2001), but this increase was much lower than the widening of the banking spread.39

As a matter of fact, the higher intermediation cost by nationalized banks, given their size in the banking industry, is creating an upward bias in the cost structure of overall banking business in Pakistan.40 It is generally believed that the large operating cost for this group is mainly attributed to their huge infected portfolio, overstaffmg and over branching, and disproportionate tax burden.

The discussion that follows endeavors to explain the dynamics of the banking spread during the last decade. It is essential to mention that the purpose of this discussion is not to identify reasons for the higher spread in a static setting, but to analyze the factors responsible for increasing it in the changing scenario. More specifically, the following sections will highlight how bad loans, advance taxes on banks, and a decreasing share of concessional lending, affected the banking spread, over time.

Over time, rising non-performing and defaulted loans usually result in increased intermediation costs for banks.41 This is because infected loans not only constrain the earning opportunities, but resulting provisionings also increase the operating expenses.42 Hence, this situation compels banks to increase the spread further to maintain profitability. As shown in Figure 3, both NPLs and defaulted loans depict a rising trend during the 1990s. It is important to note that Nationalized Commercial Banks (NCBs) were the biggest contributors to both, the stock and growth, of bad loans of the banking industry.43 This is primarily because substantial loans were provided by NCBs on political grounds. In addition, stricter disclosure requirements put in place by SBP forced banks to disclose a truer picture of accumulated bad loans, pushing up the aggregate figure. This alone resulted in considerable rise in NPLs during the 1990s, and especially in CY99. Not surprisingly, this would have added to pressure on profitability during the period, leading to larger spreads.

Another major factor that contributed towards the high spread was the ageing of bad loans that further deteriorated the balance sheets of banks.44 Despite provisioning, banks (especially NCBs) had not been able to clean their balance sheets, mainly due to poor quality of underlying collateral and possible legal complications due to lacunas in the respective judicial framework. In order to remove the legal difficulties and time delays faced by banks in recoveries against defaulted loans, the Financial Institutions (Recovery of Finance) Ordinance, 2001 has been promulgated. This will help banks in recovering the stuck up loans by the right foreclosure and sale of mortgaged property with orwithout intervention of courts.

Banking is a heavily taxed industry in Pakistan. While the tax rate on banks has been gradually brought down from 66 percent at the beginning of 1990s to 47.0 percent by July 2002,45 it is still high compared to that in regional countries, such as 40 percent in Bangladesh and 36.0 percent in Sri Lanka.

The declining tax rates during the 1990s should have helped banks to maintain after tax profits by operating at relatively lower interest rate margins. However, even though the banking and interest rate spreads (see Figure 1 and 2) rose, the profitability of the banking industry fell during second half of the 1990s (see Figure 4). The contributory factor to this dismal performance was the problem of advance deduction of taxes.

More specifically, since (1) banks were paying 30.0 percent withholding tax on T-bills46 and 10.0 percent on FIBs/PIBs, and (2) the bulk of the interbank transactions in T-bills, this created a substantial accumulation of advance tax payments.47 It is important to note that the advance taxes not only reduced the after tax profit but also increase the non-earning assets of banks, forcing the institutions to increase their spreads to maintain profitability. This tax burden, computed by dividing the total tax paid by banks during 1990s with total profit of banking system, reached as high as 83.8 percent. It is essential to mention here, that this figure is a conservative estimate; for a better approximation one should also consider the opportunity cost of such unproductive tax assets. It is also important to note that this large tax burden was primarily because of nationalized banks; the tax paid by the other groups was close to their actual tax liability.48 Realizing the severity of this problem, government had already taken a significant step by issuing PIBs worth Rs 22.0 billion to these banks against their tax adjustment receivables in June 2002. This will help not only in increasing the profitability of banks but also in lessening the spread in future.

One may recall the fact that lending in Pakistan is not fully market-based. A large amount of credit has been extended under various concessionary and mandatory schemes, depriving the banks (generally higher) market-determined rates on their funds. The presence of subsidized credit has implications for the computation of weighted average lending rates, as the resulted figures may not reflect the true market rate and have downward bias. However, the share of this concessional credit in total credit has gone down substantially from 44.2 percent at the beginning of the 1990s to 31.0 at end June 01. Moreover, the subsidy margin has also declined significantly after FY98 (see Figure 5).49 It is important to note that changes in both the share of concessional credit and margin of subsidy over time may alter the degree of downward bias in calculating the weighted average lending rates.

Case I: Share of subsidized credit declines with the same margin of subsidy during period 1 and 2 Case 2: Margin of subsidy declines with the same share of subsidized credit during period 1 and 2 Case 3: Both share and margin of subsidy decline during period 1 and 2

The declining share of subsidized credit has the tendency to push up the weighted average lending rates over a period of time, even if the market rates are not increasing.50 This would lead to the conclusion that a simple comparison of weighted average lending rates between two periods may not be advisable, as the weighted rates incorporating a larger share of subsidized credit would have a larger downward bias (see case 1 in Table 1 for hypothetical examples). Another important dimension of concessional lending schemes that may influence the average lending rates over time, is the changes in margin of subsidy. A decline in margin may reduce the downward bias in calculating the weighted average lending rates (see case 2 in Table 1).

Hence, an adjustment is required in lending rates and spread for a meaningful comparison between two different points of time. As shown in Table 2, making the adjustments by excluding subsidized credit, spread has gone up by only 0.6 percentage points,51 against 1.4 percentage point in overall lending rates (including both market-based and subsidized credit).

It is not advisable to use banking or interest rate spread alone to gauge the success of banking efficiency. A part of this increase in spread may be attributed to cost of reform process, as disclosure of true classification of loan quality and provisioning requirement under strict supervision by SBP added to the operating cost of banks. However, this has contributed positively towards the soundness

Table 2: Weighted Average Rates of Return

percent

Note: Subsidized and market rates are calculated weighted average rates on concessionary and non-concessionary lending respectively, while overall rates are on overall credit, same as published in SBP Banking Statistics.

of banking system, as earlier financial statements of banks were depicting better than their actual health. Hence, this higher spread may be treated as cost of this improvement. As discussed in Section 6.3.3, nationalized commercial banks in particular and the overall banking sector in general saw improvement in their asset quality. NPLs of the banking system are on downward trajectory since Q3-FY02 and banking spread has narrowed during FY02.

Similarly, efforts towards liberalization of interest rate structure as a part of reform process also contributed towards higher spread, earlier interest rates were controlled through floors on deposit rates and ceilings on lending rates. Moreover, decline in subsidized credit and subsidy margin overtime made the weighted average lending rate relatively closer to market rates, thus simple comparison at the beginning and end of 1990s are misleading.

Though the tax rate on the banking system was gradually brought down during last decade, but bank especially in public sector ended up paying huge amounts in the form of advance taxes. This was not only draining the after tax profit but also the future flow of income of these banks. This exogenous factor has contributed towards higher spreads and hence banks alone may not be held responsible for this. Ambitious tax revenue targets by IMF and limited option available to CBR may be a possible cause for this development. Banks saw some relief in this regard towards the end of FY02; Rs 22.0 billion PIBs have been issued to banks against their claims on CBR. This should help banks in narrowing down the spread further in coming months.