The Bangladesh Bank has made a candid disclosure of the risk to financial stability in Bangladesh in its Financial Stability Report 2022. The key message from the data presented is not a comforting takeaway when you are thinking about the present and future solvency of the banking system.
Finance is supposed to be a conjunctive intermediary facilitating economic activity through payment, safekeeping, lending, and risk management services. Bangladesh's banking system, a proven dynamic force, has been struggling in the last two roles. Asset growth (7.5%) in 2022 was the lowest since the pandemic. Growth in loans and advances (13.5%), the largest (68%) part of banking sector assets, was better. However, the quality of loans has deteriorated to an extent that the industry, at the aggregate level, is close to breaching solvency standards.
Buffers on the edge
Capital to risk-weighted assets ratio (CRAR), a measure of how much money banks have relative to their loans and other assets adjusted for the likelihood of some going sour, is too close for comfort to the minimum 10% capital requirement. Despite increasing by 75 basis points in 2022 to 11.83%, the aggregate CRAR is short of meeting the 2.5% capital conservation buffer (CCB), the extra amount of money banks must keep as a safety net to face adversities. Banks are expected to limit dividend and bonus payments when the CCB falls below the prudential limit.
The buffer is actually less than what you may construe from the data. Underlying the increase in CRAR is the increase in net profit after taxes of the banking sector in 2022 relative to 2021 from Tk50.2 billion to Tk142.3 billion respectively. This happened at a time when the growth and quality of assets deteriorated. Could it be that the reported net profits are overstated when measured by BB's regulatory standards?
The answer unfortunately is affirmative. Under provisioning against non-performing loans was equivalent to 77.3% of the reported profits in 2022. The provisioning shortfall spiked in the last two years. The maintained provision rate to gross NPL ratio decreased from 64.5% in 2021 to 60.6% in 2022. The CCB shrinks to 0.1% of the risk-weighted assets, compared with the reported 1.8%, when under provisioning is accounted for.
Thanks to BB for acknowledging its own role in inadequate provisioning: "the current level of CRAR…may not sustain if the central bank withdraws the accommodative provisioning stance…" (FSR 2022, p.59).
Accounting for the fragility
The precarity of the banking system is rooted in low asset quality which is partly out in the open and partly still in the closet. Non-performing loans (NPL), a metric of perishability of assets, increased to 8.2% in 2022 from 7.9% in 2021. These figures represent the loans recognised under the existing lenient nonperforming loan recognition policy of the BB. Recognised NPL is dominated (over 88%) by the worst category called BL (bad loans). These are loans classified as nonperforming for 12 months or more. The total BL of the banking sector increased 17.5% in 2022 relative to 2021, miles ahead of the 2.9% growth in total NPLs.
There is a window-dressed part of the NPLs which do not show up in the above numbers. These are the loans rescheduled, because they turned bad in the past, and the amounts written off because they have been on the balance sheet as BLs for more than three years.
The share of rescheduled loans in total loans increased from 12.9% in 2021 to 14.4% in 2022. This cannot be added, as some seem to be doing, to the NPL ratio because it includes the part of the rescheduled loans which have turned non performing again and hence are included in the reported NPL. A nontrivial 19.6% of rescheduled loans in 2022 were nonperforming, constituting 33.9% of recognised NPL. Adding the stock of written off (net of recovered) and the non-classified part of the rescheduled loans to the recognised NPLs, the incidence of "distressed" assets was 23.9% in 2022, equivalent to about Tk3,600 billion (1.4 times the planned FY24 Annual Development Programme). This does not include the known unknown billions of loans stuck in courts.
The aggregates hide important details
The culprits within the banking system are a heavyweight minority that is no longer the exclusive domain of the public sector. The asset quality of some significant private commercial banks (PCBs) have started to look more like those of the state owned commercial banks (SOCBs).
The top five and top 10 banks accounted for about 46% and 65% of total NPLs respectively in 2022, rising from 43.5% and 62.5% in 2021. The top 10 in 2022 consisted of five SOCBs, four PCBs, and one specialised development bank (SDB) based on NPL size. The provisioning shortfall is concentrated among four PCBs, three SOCBs, and one SDB. The provisioning shortfall of SOCBs declined but the provisioning surplus of PCBs in 2021 turned into a shortfall in 2022, driven by about a handful few.
Asset concentrations within the top five and top ten banks were 30.7% and 44.6% respectively in 2022, decreasing from 31.6% and 45.6% in 2021. The top 10 banks comprise six PCBs and four SOCBs of whom several are already insolvent. Sectoral loan concentration increased in 2022 in large industries, wholesale and retail trade, and import financing. Banks engage more in corporate loans as evidenced by the 26.5% market share of large industries in total loans.
NPL concentration mirrors loan concentration and more. The good news is out of 61 banks, all but one foreign commercial bank (FCBs) and all the PCBs except five recorded single-digit gross NPL ratios in 2022 with 38 maintaining it below 5%. The bad news is the few (three SOCBs, three PCBs, one FCB, and one SDB), who had NPL ratios over 20%, are not all small players. The top five and 10 banks had about 46% and 65% of aggregate NPLs respectively. The fastest growing manufacturing sector accounted for 49% of the loans extended and 55% of NPLs. Such disproportionalities are particularly notable in cases of RMG, textiles, ship building and ship breaking.
Most of the rescheduled loans were in the industrial sector. Ship building and ship breaking topped the rescheduled list with 33.8% share in 2022 followed by RMG and textile. Out of total outstanding rescheduled loans, the top five banks held 38.3% while the share of the top 10 was 58.4%. The top five banks were composed of three PCBs and two SOCBs and the top 10 consisted of three SOCBs, six PCBs, and one SDB.
There is likely to be a significant overlap between the concentration of loans, the propensity to default, and the bank types (ownership, generation). While SOCBs are a common denominator in all clusters with a significant share of assets and even more significant share of NPLs, some PCBs have metamorphosed to the SOCB like balance sheets. The persistence and spread of this phenomenon defies technocratic explanation. A story centring around the symbiosis of interests in a dynamic contest between market and political power holds better promise.
A bedsore if not a ticking time bomb
BB's stress test correctly identifies the dominance of credit risk as the most material and plausible source of risks to the banking system. A 3% increase in NPLs pushes capital adequacy down below the regulatory minimum. If in fact we allow for increased provisioning against the unclassified rescheduled loans, breaching the regulatory minimum could happen with as low as a 1.5% increase in NPL.
The return on equity did not increase enough to stem the slippage towards breaching the regulatory minimum capital buffer. Reported ROEs increased from 4.7% in 2021 to 10.7% in 2022. Is this good enough? Given 8% growth in total financial assets, 3.6% growth in risk weighted assets, 4% growth in required capital, as it were in 2022, the ROE required to comply with the regulatory minimum CRAR is 16.1%. Dividends paid when the actual ROE is below this point, as was the case in 2022, eroded capital. These are not manifest in the accounts because of under provisioning.
In the event of a combined shock (NPL, market risk, exchange rate risk and interest rate risk), despite excluding defaults by top large borrowers and an increase in NPLs in the highest outstanding sector, the banking system's CRAR declines to 7.4%, according to the FSR 2022. Assessing the likelihood of scenarios that could produce the combined shock equivalent or even larger effects (if default by top three borrowers are included for instance) is important to pre-empt any disruptive trigger in the economy where banking is the epicentre.
There is liquidity crunch already in a number of banks ignited by high NPLs, weak deposits growth or even net cash outflow in some banks, fraud and embezzlement of funds and so on. With at least 11 banks undercapitalised for many years, additional five Islamic banks intervened by BB in November-December 2022 by appointing observers and providing a special liquidity support window, the combined shock equivalents no longer look as hypothetical as they used to.
Do not wait for a crisis to force reforms
We are likely to see more banks in a similar situation in the next report if the authorities continue on the path of business as usual. Risk aversion is the natural recourse of prudent players in a system littered with moral hazards. Axing bad apples when the bad ones are not readily ostensible inevitably axes good apples as well. Growth and inclusion are thus affected as innovations, startups, and entrepreneurs without tradable collateral get excluded.
Bad assets in a banking system tend to be toxic. The use of forbearance and directives in diverse forms for more than a decade now to improve asset quality has not delivered results. It may have created handsome rents for lawyers, accountants, auditors, and regulatory supervisors. Recent legislative changes are no game changers in the needed direction. Even muddling forward becomes difficult as the counterparties and third parties react to the reputation loss caused by the persistence of systemic malware.