Book excerpt: How RBI under Raghuram Rajan started the war against NPAs

The Reserve Bank of India (RBI) could sense that something was wrong but it was helpless as it did not possess the full picture. Until the establishment of the Central Repository of Information on Large Credits (CRILC) in June 2014, data on NPAs were disaggregated.
 
The CRILC was Rajan’s initiative. Once it was in place, the banks started supplying data in real time for all loans of Rs 5 crore and above. For accounts turning bad, the data was weekly, while standard accounts were reported monthly.
 
Banks had to submit quarterly reports on all borrowers with an aggregate fund-based and non-fund-based exposure of Rs 5 crore or more. They also had to classify borrowers as Special Mention Accounts (SMA) of various levels to gauge the probability of such accounts turning bad…
 

Apart from regular quarterly submissions, the banks were advised to keep the regulator informed on a real-time basis whenever a large borrower's account becomes overdue for 61 days (SMA-2) and/or banks were jointly planning to restructure such an account. The special mention accounts and the step by step classification of non-payment dealt a blow to the cosy relationships between banks and borrowers.
 
The RBI, as well as banks, could access this data. This gave a comprehensive view of the banking system’s exposure to every large borrower and how the exposure to the same borrower was classified differently by different banks. Most importantly, the central bank could see how funds were moved across banks to keep accounts ‘standard’.
 
It was clear that the situation was far worse than what the RBI had envisaged. In its worst nightmare, the regulator could not have imagined the terrible asset quality…
 
Risk-Based Supervision
 
Just a year before CRILC came into existence, the RBI had changed its supervisory model – from CAMELS (capital adequacy, asset quality, management, earnings, liquidity and system and control) to RBS or risk-based supervision…
 
Once this was done, RBI could see that even smaller banks had relatively large exposures to corporate accounts. Growth-hungry banks depended on the lead bank (in most cases, the SBI) for project appraisal and gave loans to large corporations even though their balance sheets could not justify such risk-taking. Technically, it is called multiple lending but, for all practical purposes, this was informal loan syndication.
 
No bank wanted to miss out the credit growth story. Between 2006 and 2009, bank credit grew at a scorching pace. It dipped after the Lehman crisis but started to pick up again in 2011.
 
Another factor that contributed to this trend was the memorandum of understanding public sector banks (PSBs) were signing with the government every year for their performance. The CMDs and executive directors were given a bonus every year. This was a pittance but it was linked to balance sheet growth and not the quality of assets.
 
In many cases, the banks appeared to be drafting the proposals as well as appraising them. The large banks, the bullying big brothers, were calling the shots in making decisions on which borrower to finance and by how much. The smaller banks were forced to fall in line.
 
The RBS opened a new can of worms.
 
The RBI discovered the following:
 

• The default risk in almost all banks was extremely high. Typically, the default risk is measured in standard asset portfolio of the banks over and above the declared NPAs.

• The recovery risk was also equally high because enough security had not been taken when loans were sanctioned. In most PSBs and some private banks, it was usual for the higher officials to permit deviations in security norms.

• The stressed assets were relatively higher in industry and lower in agriculture.

• Certain banks, particularly a few private banks, had efficient managers and their quality of loans was better.
• The low-capitalised banks tended to take more risks for more returns as higher returns could help them build capital.

• Highly profitable banks, too, tended to take more risks for better returns.

• Banks with a diversified portfolio of retail and corporate loans had better balance sheets.

• There were issues with sectoral and geographical concentration – such as real estate, steel, tea gardens in Assam, among others.

• There were differences in the quality and quantity of collaterals held by different banks.

• There was rampant misuse of the general-purpose loans to companies for business. The banks were giving such loans to tide over liquidity crisis and thus, keep accounts performing.

Once the RBS was in place, most banks were downgraded at least by one notch and many even lower, subsequently.
 
The regulator’s concerns were on multiple grounds:
 

• Poor credit assessment of banks.

• Lack of monitoring of loan accounts.

• The banks did not assess the high leverage of corporate borrowers.

• The borrowers were siphoning off money and the banks could not track that.

• The promoters’ contribution and personal guarantee for loans were only on paper; in most cases, they were not charged to banks and hence unenforceable.

• The banks were restructuring the loans by giving fresh loans but the promoters were not bringing in any contribution in the form of equity.

• There was a concentration of risk as all banks were rushing to lend to certain sectors such as steel and infrastructure, without the skills required to appraise projects in those areas.

• The banks were putting in a lot of conditions before disbursing a loan but waiving many terms later. For instance, a large steelmaker threw in many pieces of prime real estate in Mumbai as collateral. But later it convinced the banks to release the real estate as the project itself (land, plant and machinery) was hypothecated to the banks. The problem was that the land of such projects could not fetch much money and the plant and machinery turned into scrap.

 
Were the banks compromised? The RBI did not explicitly say so. But it found that, after the sanction of loans, if the supervision was bad, monitoring was worse. Since the project appraisal was done by the lead bank, others were relaxed. Also, the bankers were taking borrowers’ commitments at face value.
 
Another ingenious system of managing bad loans was discovered. When a string of accounts was on the verge of turning bad, the bank typically created a pool of debt covering all, got it rated by a not-so-reputable rating agency, and sold it to relatively smaller banks, including cooperative banks. This practice of down-selling or palming off of bad loans was invented by a private bank, which later came close to collapse.
 
Yet another clever way of managing bad assets was sanctioning two loans at the same time. While the first loan would be disbursed immediately, the disbursement of the second loan was typically kept on hold and, the second loan got disbursed to help the borrower if the first loan turned bad!
 
CRILC Review
 
By November 2014, Rajan called for a review of CRILC, which had been in operation for six months by then. The focus was on how banks were reporting their bad loans. Every account worth Rs100 crore or more was scrutinised.
 
By that time, the banks had stopped ‘restructuring’ bad loans but ‘rectification’ was being done. To its horror, the RBI found that rectification had replaced restructuring and most large accounts were rectified. How could every large borrower have a cash crunch?
 
Around the same time, at a workshop on the new supervisory model, the RBI assessed the progress of the risk-based supervision. The Department of
 
Banking Supervision (DBS) of the central bank decided that a thematic review should be undertaken on important aspects of credit, market and operational risks.
 
To start with, it was decided that the asset quality of the big banks be simultaneously reviewed. In March 2015, the process was complete and Rajan was informed.
 
The Cat out of the Bag
 
Around half-a-dozen large banks including SBI, PNB, ICICI Bank and Axis Bank were asked to make presentations to the RBI. The regulator got a clear sense of what was happening. The cat was out of the bag.
 
Deputy Governor S.S. Mundra and Governor Rajan decided to take a closer look. They put the largest 100 borrowing accounts in each bank under the scanner.
 
What started with a basket of 100 large accounts expanded manifold. What were small accounts for say a large bank such as SBI were large accounts for a smaller bank such as Dena Bank, or Bank of Maharashtra. Hence, even with the focus on 100 large accounts of each bank, the RBI ended up scrutinising thousands of accounts.
 
Rajan and Mundra discovered accounts which were registered as NPAs in one bank were supposedly performing assets in another. Many accounts were not restructured but these had been rectified repeatedly.
 
Rajan’s patience was running out. Enough time had been given to the banks for restructuring stressed loans but they were misusing it, and trying to take the regulator for a ride.
 
First of its Kind
 
That exercise by Mundra–Rajan unearthed so many problems that the RBI decided that a system-wide audit was needed. This is how the asset quality review or AQR – a first-of-its-kind health check of Indian banking – was conceived.
 
Globally too, there was nothing comparable to the AQR anywhere. After the collapse of Lehman Brothers, the US Federal Reserve started conducting sensitivity test and solvency tests to identify bank which could become basket cases but the banking tuft in the US is very different.
 
Most US banks have exposure to market instruments and this makes it easier to flag problems. If a bond is issued and not serviced, there’s an issue. But Indian banks are primarily in the business of giving loans. Loans work via agreements between borrower and bank, and banks can accommodate borrowers in many ways.
 
Typically, the annual RBI inspection starts after July when the banks have released their audited results; this is the second quarter of the Indian financial year. The inspection continues through the financial year until the end in March of the next calendar year. Depending on the seriousness of the findings of inspection teams, interactions with the bankers can continue for months.
 
A simultaneous look at all banks through the AQR is a very different story.
 
Anxious Bankers
 
During financial year 2014–15, bankers were looking anxious. They were ‘negotiating’ for regulatory forbearance. The RBI could sense the unease in the same way that a doctor feels there’s something wrong about a patient.
 
Everyone agreed that something had to be done. But when and how? Who will bell the cat? Did the banks have the capital to absorb the shock of new NPAs?
 
The AQR involved the focussed direction of supervisory resources. It would throw new light upon risks within the system. The DBS planned the scope, modalities and execution meticulously. An AQR team was constituted with the governor as its chairman, and all the proposals of the DBS were discussed by this team.
 
The Go-ahead
 
The AQR process started in June 2015, after Rajan gave the go-ahead. It was a three-month project, slated to be completed by August 2015 so that the recognition of bad assets could start from the September quarter. But it spilled over to October.
 
The clean-up process started from the October–December 2015 quarter.
 
This is how it worked.
 
India has over 100 banks. But very small ones, particularly foreign banks with small balance sheets and no corporate accounts, were left out of the exercise. The RBI used all the resources at its disposal to make a success of the AQR.
 
A team was formed for each bank, and all teams worked simultaneously.
 
Typically, a senior supervisory manager (SSM) looks after the audit of a bank. The rank of the SSM depends on the bank being audited. An SSM supervising SBI could be a general manager at RBI, but the SSM of a relatively smaller bank could be a deputy general manager or even as assistant general manager. For large banks, a team of four executives can work under the SSM.
 
The Control Room
 
The AQR team was like a financial bomb-disposal squad, trying to defuse explosives which could destroy the depositors’ trust – the bedrock of banking.
 
A control room was set up on the third floor of RBI’s Cuffe Parade office, away from its central office on Mint Road. Two general managers were collating the real time data; a third general manager was consolidating them under different heads; and a fourth one was looking at the big picture on the computer screen.
 
The chief general manager in charge of DBS was heading the control room. For six months, the team hardly slept. All of them worked flat out, even over the weekends.
 
Every evening the Cuffe Parade team would come over to the central office to take stock of the situation. Over poha and coffee from the canteen, the meetings would carry on at least till 10 pm. But quite often, there would be calls at midnight, or later, from senior colleagues who stayed tuned in 24x7.
 
Slowly, a big picture emerged. Typically the team discovered anomalies. Account ‘A’ was a performing asset on the book of Bank X but non-performing asset on Bank Y’s book and a stressed asset on the book of Bank Z.
 
 How could that happen?
 
Well, there are many ways:
 

• Routing the same credit through multiple banks

• Raising temporary finance to service debt from other private and foreign banks under multiple bank finance.

• Switching to different finance models.

• Creative products such as capex loans – an all-purpose, no-appraisal sort of loan repayable at the borrowers’ convenience with incredibly long tenures and full repayment only on maturity.

• Hiking the borrowers’ cash credit limit, allowing them to draw more money to pay back loans.

• Misuse of export guarantee covers.

• When Letters of Credit (LCs) were devolving, banks were not recognising devolvement.

• Giving short-term overdrafts to borrowers to repay loans.

• Offering fresh loans at the next stage to adjust the overdraft.

• Artificially raising the sales volumes to lift the limit of bank borrowing.

• Blatantly misusing the CDR platform for restructuring.

• Seeking personal guarantees of the promoters but not invoking them.

• Arbitraging between fund-based and non-fund based exposure by manipulating respective limits.

• Most of the technical evaluation studies were rotten. While the borrowers were faltering on every business parameter, the studies were justifying higher credit limits.

•  The date of commencement of commercial operation (DCCO), a critical milestone for any project loan, was manipulated. Often, there was a date on which a factory ceremonially opened. But there was no production. It was just a formality.

 
The DCCO is the date when the unit is supposed to begin operations. The sales and other financial projections, including the repayment schedules for the loans, are based on this date.
 
The more the date was postponed, the longer the tenure of the loan and the bigger the interest liability and hence the higher the probability of the loan going bad. The restructuring norms permitted postponements of DCCOs in certain circumstances, but this was rampantly misused.
 
The RBI found that the banks were delaying recognising large accounts as NPAs, even if projects were not starting up, and the ‘project commencement’ milestone was largely cosmetic.
 
In one such case, the RBI team found that a huge sum of money had been deployed overseas for oil and gas exploration. But just an Internet search revealed that the oil field the banks had funded in Africa had already been identified as dry by a major oil company.
 
This prompted the regulator to launch ‘targeted scrutiny’ of certain accounts and ‘thematic scrutiny’ of real estate, housing loans, bills and letter of credit discounting.
 
It also went for ‘targeted audits’ of certain accounts by external auditors to red- flag misdoings. The banks themselves were appointing such auditors, at the instance of the regulator.
 
House of Cards
 
Given the scale at which the system was being gamed, the RBI had to opt for carpet bombing.
 
Among other things, it found innovative usage of the so-called standby letter of credit. This is a guarantee by one bank, based on which another bank would lend. The RBI found these were used as chain letters – Bank A was giving money, based on a guarantee by Bank B, then it was passed on Bank C, D, E, F and so on.
 
The same bank was becoming the guarantor as well as the lender. This was a house of cards but no bank involved in this chain would ever invoke a guarantee, since that would bring the house down. The borrowers had no skin in the game; the bankers were funding both debt and equity.
 
Instead of paying banks’ money from their own pockets, the borrowers were also using accommodation bills to service the principal plus interest of previous bills. Backed by the new bill, they would get fresh and higher credit, which would enable them to clear the previous loans and the chain continued.
 
As if this was not enough, there was also round-tripping.
 
Round Tripping
 
One of the SSMs found a large account in a very large bank, let’s call it Bank A, had not been serviced on the 90th day. Why was Bank A not classifying this loan as an NPA? The SSM was informed the required funds would flow into its account before the end of the 90th day. Indeed, the money arrived and the SSM was impressed – here was a banker who did not believe in hiding facts.
 
There was no need to classify that account as an NPA. So, it was maintained as a standard asset. But just out of curiosity, the SSM checked the CRILC data the very next day.
 
Lo and behold, to his horror, he found an equivalent amount of money had left Bank A and gone to another bank, Bank B. He couldn’t believe his eyes. But one phone call to his counterpart in Bank B confirmed his suspicions. But the story did not end there! The funds left Bank B a day later and entered the books of a third bank, Bank C.
 
What was going on?
 
These banks were keeping the account standard simply by transferring money from one bank to another, to yet another. In such chains, as many as ten banks might have sanctioned a loan to one borrower. But the dates of sanctions and disbursements were different, allowing for the round-tripping. It may be inferred that loans were applied for, and sanctioned, on different dates deliberately to enable this sidestep of the NPA recognition norms.
 
The Next Step
 
 Broadly, the AQR findings can be divided into four categories:

• Banks were piling up NPAs but not recognising them.

• They were camouflaging NPAs through fresh loans and other innovative ways.

• The CDR platform for restructuring stressed assets had failed.

• The date of commencement was being gamed. Many projects had not got going on the formal commencement dates.

 
There were long discussions about the next step. Deputy governor Mundra was conservative in his approach and wanted to give more time to banks. But his boss Rajan thought the rot was too deep. He could not accept the manipulation of the DCCO.
 
What next? It was decided to analyse all the findings once again, in stages.
 
At stage one, the focus was on what they call ‘quality assurance’. Here, three CGMs and the CGM in-charge of DBS analysed the tabulated data. They followed up with a series of meetings with the group of SSMs to get a clearer picture before confirming each NPA account.
 
Once the consolidated picture emerged, at stage two, a presentation was made to Mundra, Rajan and the executive director in charge of supervision, Meena Hemchandra.
 
At stage three, the RBI supervision team met the bankers one-on-one and discussed the findings. The venue was the Cuffe Parade office, which resembled a sort of Scotland Yard during these meetings. The bankers were not grilled but the
 
RBI team had every detail on the table to confront them. There was documentary evidence to support every point they were making.
 
After the meeting with the bankers, at stage four, a second round of checks was undertaken, with the bankers’ feedback and more facts.
 
At stage five, it was decided that 30 of the largest accounts must undergo a second level of quality assurance check. This was done by senior executives of the supervision team who had not been involved in the exercise. They were flown in from Delhi and Chennai for a third-party check.
 
At stage six, the control room did an assessment of the impact. If the banks were to recognise all bad loans at one go and set aside money, how would it affect their NPA pile, capital and profitability? The team looked at the last three years’ profits, presumed a 5 per cent rise each year and worked out different scenarios on the assessed NPA figures to estimate the impact.
 
Once the exercise was over, the team went back to Rajan to make the final presentation, drawing up different scenarios about the capital requirement for public sector banks.
 
Before biting the bullet, at the last stage, Rajan called a meeting of the bankers at the RBI central office on Mint Road. All big banks present there were vociferous in their protests. The boss of a large bank said the RBI was not following the rules, and the decision to force the banks to clean up the bad loans was principle-based.
 
A rules-based approach to regulation prescribes in detail a set of rules on how to behave as opposed to a principle-based approach to regulation where outcomes and principles are set and the controls, measures and procedures on how to achieve that outcome is left for each organisation to determine.
 
Is this True?
 
In his typical style, Rajan asked the team one simple question: ‘Is this true?’
 
The 15th floor conference room at the central office has seldom seen a debate as animated as the one that occurred. The team explained graphically to the governor how every rule was followed and nothing was left to principles.
 
It backed up this argument with multi-level evidence. The bankers had no choice but to accept what the RBI team was saying.
 
Immediately after the bankers’ meeting, Rajan spoke to the junior minister of finance, Jayant Sinha, who had been his classmate at IIT Delhi, on a video conference.
 
Sensing the imminent collapse of all the creative scheming, borrowers and bankers started lobbying in Delhi to stop the inevitable. They sought the shoulders of influential ministers to cry on.
 
In November, Rajan, along with Mundra and the chief general manager in charge of DBS, flew to Delhi to meet finance minister, Arun Jaitley. The mission was to convince Jaitley that the clean-up exercise could not be avoided.
 
The finance minister gave them a patient hearing, got a sense of the capital requirement for the banks and gave the go-ahead. His junior minister Sinha and Anjuly Chib Duggal, the financial services secretary, were present at the meeting. The next day, Jaitley announced the government was ready to clean up the banking system. It came up for discussion next month, on 11 December, at RBI’s 555th Central Board meeting in Kolkata.
 
Excerpted from Pandemonium: The Great Indian Banking Tragedy by Tamal Bandyopadhyay, courtesy Roli Books. Releasing on 9 November 2020, Price Rs 695.