FLASHNEWS:

JS Securities Limited – JS Research (30 -03 -2023)

Karachi, March 30, 2023 (PPI-OT): Banks: Dissecting CY22 loan growth and asset quality

In the midst of a sharp monetary tightening cycle and macro challenges, banking sector regulations were revised with the objective of discouraging lower ADR. This led to banks cherry-picking deposits in CY22, resulting in 14-year low deposit growth at 7% YoY while expanding loan book with double-digit growth in order to comply with normalized tax rates.

We analyze the segment-wise loan growth across banks where we observe loan growth was concentrated to limited segments, while most common growing segments were Textiles, Individuals (Consumers) and Financials.

The worrisome part – Textiles, emerged amongst the highest contributors to asset quality deterioration as well. While loan contribution remains at ~17%, challenges faced by the sector add concerns to prospective credit cost from the segment.

With implementation of IFRS 9, banks measuring loss expectations on investments and loans by assigning probability on forward-looking scenarios, the course of credit cost computation would witness a major change. Coupled with macro challenges, we keep our credit cost estimates at 100bp annually for CY23E and CY24F.

Pressure to increase ADR in the midst of macro challenges

A monetary tightening cycle generally invites concerns over asset quality of banks and potentially large provisioning expenses. This cycle, that has witnessed a 1,300 bp expansion in Policy Rate (20%) in past eighteen months, is no different when it comes to similar concerns.

Moreover, in midst of these concerns, the sector faced regulations discouraging a lower ADR by penalizing the same via higher taxes. This led to banks cherry- picking deposits in CY22, resulting in 14-year low deposit growth at 7% YoY while expanding loan book with double-digit growth in order to comply with normalized tax rates.

Textiles, Consumer and Financial segments appeared popular

We analyze segment-wise loan growth of various banks where the most common growth segments to emerge include Textiles, Individuals (Consumers) and Financials. It was also noticed that the loan expansion was relatively concentrated with the top three growth segments contributing on average 67% of total growth (ranging from 50% to 100% of the growth for individual banks).

This keeps textiles amongst the larger contributing segments in the banking sector’s lending portfolio for our sample base (~17% of loan portfolio). In addition, while the Consumer segment also witnesses a larger increase, its weight remains relatively lower at 9%, similar to the weight maintained since 2016.

The recent data does bring some flash back from the sour tasted 2008 financial crisis. To recall, the crisis led to hefty NPLs fuelled by higher Advances to Deposit levels of c. 80%, lending towards high-risk sectors and expanding leverage ratios of corporates. The increase of 500bp in Policy Rate during 2008 led to the banking sector’s NPLs to expand by almost two folds by 2011. This led to infection ratio increasing to an alarming level of 16%, from 8% in 2007.

Over the years, the banking sector’s Infection ratio has improved by limiting lending to high quality assets and diversifying lending to other sectors. With loan growth opportunity arising from the Energy sector over the past couple of years, the banking sector has increased share of Production/Transmission of Energy segment in its loan-book, leading to partially substitute the share of Textile segment in the total pie.

Asset quality remains in check

Analysing the NPL stock increase in CY22, though absolute stock has increased by Rs40bn, a quantum last witnessed during COVID – broadly related to one segment – and before that in 2010. The infection ratio has however improved by 80bp YoY to 5.8% in CY22. The contribution to NPL accretion has been a mixed bag with various segments from some banks while concentrated for others, however, Textiles again being a segment highlighting more than others. A few banks were also observed to avail a higher Forced Sale Value benefit when compared to their historical trend.

Though among our covered space, Coverage Ratios for seven banks surpass the 100% mark, while sample average comes to 100%, we expect worsening asset quality to stem from the consumer, partially textiles and other miscellaneous segments.

Higher exposure to textile may increase concerns

The worrisome part – Textile sector, emerged amongst the highest contributors to the asset quality deterioration as well. In 2008 financial crisis, Textile sector contributed c. 20% to total loan portfolio of the banking sector, with a high gross infection ratio of 29%.

Catering to robust global demand post COVID, the Textile Sector of Pakistan witnessed an impressive turnaround with the sector’s exports touching a historical high of US$19bn, up 25% YoY during FY22. The growth momentum has, however, tapered off as global economies including key export destinations such as US and Europe move closer towards a looming recession. The shift comes as economies across the globe tighten their monetary stance in an attempt to battle soaring inflation.

Impact has been felt in the local industry with recent data reporting Textile exports declining by 30% YoY, the impact as per industry stakeholders is expected to accelerate in the coming months with pressure on margins. The industry faces a number of obstacles to growth such as shortage of gas in the winter months, destruction of cotton due to floods and delays in sales tax refunds from the GoP leading to a severe liquidity crunch, despite some benefit of PKR devaluation on the sector’s topline. In addition, gas price increase across the board worsens the case for textiles sector.

IFRS 9 implementation to change the course of credit cost

With implementation of IFRS 9 from this year, banks will measure loss expectations on investments and loans and classify it under different categories, assigning a probability to compute expected credit loss (ECL) based on developing forward- looking scenarios. Banks would be required to build and use: 1. Probability of Default (PD), 2. Loss Given Default (LGD) and 3. Exposure At Default (EAD) models based on forward-looking assumptions. These would also be adjusted to evolving economic factors.

Banks may also need to assess and assign ratings to sectors and obligors separately, which will be the basis of their working. Banks would create three categories under the name of ‘Stages’ and would place assets from least to most risky ranking from Stage 1 to Stage 3, where existing NPL stock would directly fall under ‘Stage 3’. Also, provisions under Stage 3 would be higher of as per ECL model and PR requirements. As these ECL models would be based on each banks’ own workings, data etc., the treatment of a same party may also result in different credit costs on each bank’s books.

Maintain higher credit cost estimates in medium term

In light of the above, the course of credit cost computation would witness a major change. Coupled with macro challenges, we keep our credit cost estimates at 100bp annually for CY23E and CY24F. Accounting for multiple interest rate cycles since 2007, the average credit cost of the last 15 years computes to 110bp per annum, while excluding the 2008-2010 phase of 378bp per annum, the average declines to 48bp per annum.