Viral V Acharya: Some ways to decisively resolve bank stressed assets

Speech by Dr Viral V Acharya, Deputy Governor of the Reserve Bank of India, at the Indian Banks' Association Banking Technology Conference, Mumbai, 21 February 2017.

The views expressed in this speech are those of the speaker and not the view of the BIS.

Central bank speech  | 
23 February 2017
PDF version
(58kb)
 |  6 pages

I am grateful to my colleagues and teams at the Reserve Bank of India for many stimulating discussions and insights, to many banking sector stalwarts of today and the past, and to several practitioners and policy-makers in sectors and institutions related to the resolution of bank stressed assets. All errors remain my own.

Thank you to organizers of Indian Banks' Association for inviting me today. To all the Award winners, congratulations on your stellar achievements and the innovative spirit you have shown to be today's worthy recipients. Technology, "Fin Tech" as you say, is taking banking and intermediation into unchartered territories. I have yet to fathom the full import of these sweeping changes. In the meantime, my subject matter today will be more prosaic and somewhat sobering.

I wish to speak today, with a certain sense of urgency, about the need and possible ways to decisively resolve Indian banks' stressed assets.

Since the Reserve Bank of India (RBI) initiated the Asset Quality Review of banks in the second half of 2015, it appears that possibly up to a sixth of public sector banks' gross advances are stressed (non-performing, restructured or written-off), and a significant majority of these are in fact non-performing assets (NPAs). For banks in the worst shape, the share of assets under stress has approached or exceeded 20%. This estimate of stressed assets has doubled from 2013 in terms of what had been recognized by banks. The doubling of stressed assets is the case also for private sector banks, but their ratio of stressed assets to gross advances is far lower and their capitalization levels far greater. There have been several hints - in the declining price-to-book ratios of bank equity, as I had observed in an oped piece for the Mint in September 2013, and in the incisive research reports of banking sector analysts - that many assets "parked" by banks under the Corporate Debt Restructuring cell were severely stressed. These assets were deserving of advance capital provisioning against future recognition as NPAs.

The Asset Quality Review has taken a massive stride forward in bringing the scale of this problem out in the open and stirring a public debate about it. However, relatively little has been achieved in resolving the underlying assets to which banks had lent. Several resolution mechanisms and frameworks have been offered by the RBI to banks to get this going, but the progress has been painfully slow. Most of the assets remain laden with such high levels of bank debt that their interest coverage ratio is lower than one; they have little or no capacity to raise funding for working capital and capital expenditures, or to attract private investors to turn them around. Original promoters - who rarely put in any financing and primarily provide sweat equity - have had somewhat of a field day, facing limited dilution, if any, of their initial stakes nor much of a threat of being outright replaced.

There is a connection between these two outcomes - the lack of a comprehensive recognition of stressed assets by banks and the absence of any resolution. Both stem from the structure of incentives at our banks and the fact that stressed assets have been an outcome of excessive bank lending, en masse, in a relatively short period from 2009 to 2012, and to a concentrated set of large firms in a number of sectors such as infrastructure, power, telecom, metals (iron and steel, in particular), engineering-procurement-construction (EPC), and textiles.

Let me first discuss the bank incentives. Only a bank that fears losing its deposit base or incurring the wrath of its shareholders is likely to recognize losses in a timely manner. In many of our banks, such market discipline is simply not present at the moment. In others, even if some such discipline is at work, banker horizon is excessively short until end of the CEO's term. Banks lobby for regulatory forbearance; perhaps some loan prospects have turned sour due to bad luck, but beyond a point, concessions in recognizing losses just ends up being a strategy of kicking the can down the road and leaving them as legacy assets for the next management team to deal with.

The sectoral concentration of losses substantially amplifies this problem. Given the scale of assets that needs restructuring, it is natural that the turnaround capital at asset-restructuring companies (ARCs) has been limited in comparison. Some capital is simply sitting on the fence until serious asset restructuring picks up speed. In the meantime, any assets put up for sale can raise financing only at steep discounts, implying significant haircuts for bank debt. The loss of capital that would result on bank books and the fear of vigilance actions that such haircuts might trigger have made it almost impossible to get banks to embrace restructuring.

Effectively, there is no right price at which the market for stressed assets clears if left alone to private forces. Even with an orderly resolution mechanism such as the Insolvency and Bankruptcy Code in place, why would banks rush to file cases? In the unlikely scenario that assets are in fact being sold by banks to investors at steep discounts, ARCs may just asset-strip rather than do the economic turnaround. After all, these investors have waited far too long and now wish to generate quick returns to meet the expectations of their own investors.

All this is playing out to near perfection in our setting. Its consequences are pernicious.

At one end, public sector banks are running balance-sheets that seem to be in a perennial need of recapitalization from its principal owner, i.e., the government, and shying away from lending to potentially healthier industrial credits. Bank credit growth has been steadily declining at the stressed banks. Some private sector banks face such headwinds too.

At the other end, sectors with the most stressed assets have excess capacity relative to current or near-term utilization and no sight of immediate pickup in economic prospects. Promoters have continued to operate, staying afloat with rollovers from banks which only increase indebtedness, partly disengaged, partly disgorging cash from the few assets that are running.

The end result has been a silent atrophy of the true potential of these assets.

This situation should be a cause for concern to all of us. It is reminiscent of weak banks and stagnating growth witnessed by Japan in the 1990s, with repercussions to date, and by Italy since 2010. Japan has experienced, and Italy, is in my opinion experiencing, a lost decade.

I believe we are at crossroads and have an important choice to make.

We can choose status quo, but this would be insanity, "doing the same thing over and over again and expecting different results," as Albert Einstein put it. It would risk a Japanese or an Italian style outcome.

Or we can choose to call a spade a spade as Scandinavia did to resolve its banking problems in the early 90's and the United States did from October 2008 to June 2009, even if only after letting a significant bank fail. Ireland and Spain, where the recoveries since the global financial crisis have not been as salubrious as in the United States, have nevertheless fared better than Italy; they too first adopted measures to pretend and extend troubled bank assets, but eventually recognized the scale of the problem and dealt with them in a decisive manner.

With our healthy current level of growth and future potential, with our hard-fought macroeconomic stability, with our youth climbing echelons of entrepreneurial success day after day, with our vast expanses of rural India that need infrastructure and modernization, and with our levels of poverty that have steadily declined but still need substantial reduction, we simply don't as a society have any excuse or moral liberty to let the banking sector wounds fester and result in amputation of healthier parts of the economy.

How do we embark on a better path? I have been thinking hard of ways to swiftly resolve bank stressed assets. I have tried to draw on the analysis and documentation of similar episodes in economic history that I just alluded to, which in some cases I have had the good fortune to contribute to and learn from.

Let me mention the key principles to successful restructuring that I have managed to glean:

First, there has to be an incentive provided to banks to get on with it and restructure the stressed assets at a price that clears the market for these assets. If they don't do it in a timely manner, then the alternative should be costlier in terms of the price they receive.

Second, the ultimate focus of restructuring and of assessment as to whether the restructuring package being offered to the bank is at the "right" price must be the efficiency and viability of the restructured asset. Generating the best price for the bank at all costs may only result in cosmetic changes and risk serial non-performance of the assets.

Third, not all of the resulting bank losses should simply be footed by the government. As a majority shareholder of public sector banks, the government runs the risk of ending up paying for it all. It should manage the process at the outset to avoid that outcome. Wherever possible, private shareholders of banks should also be asked to chip in. Some surgical restructuring should be undertaken to consolidate and strengthen bank balance-sheets so that private capital will come in at better valuations. It might have to accept that it is best to let some banks shrink over time. Divestments should also be on the table. Historically, significant restructuring of stressed assets has almost always involved significant bank restructuring.

Let me now elaborate "a" plan that employs two different models for stressed assets resolution and recognizes the concomitant need for bank resolution. I will provide the plan in detail to make the point that it can be done. What I enunciate should be viewed as an attempt to address all dimensions of the problem. Its individual parts are, however, not meant to be cherry-picked by or for the constituency favored by it. That would not work well.

So here it is.

1. Model I: Private Asset Management Company (PAMC). This plan would be suitable for sectors where the stress is such that assets are likely to have economic value in the short run, with moderate levels of debt forgiveness. I conjecture based on anecdotal observations that sectors such as Metals, EPC, Telecom, and Textiles, qualify for this.

  1. In terms of timeline, the banking sector will be asked to resolve and restructure, say its 50 largest stressed exposures in these sectors, by December 31, 2017. The rest can follow a similar plan in six months thereafter.

  2. For each asset, turnaround specialists and private investors - other than affiliates of banks exposed to the asset - will be called upon to propose several resolution plans. Each resolution plan will lay out sustainable debt and debt-for-equity conversions for banks to facilitate the issuance of new equity and possibly some new debt to fund the investment needs. We may have to consider that the sustainable portion of bank debt does not have to be greater than some minimum amount, so as to allow for a large haircut if necessary for economic recovery of the asset. Each plan would lay out cash flow prospects, whether the promoter stays or not, and if yes, with what stake.

  3. Each resolution plan would then get vetted and rated by at least two credit rating agencies to assess the financial health (interest coverage ratio, leverage, etc.), economic health (sector, margins, etc.), and management quality (promoter or the new team). The rating would be for the asset and not just for bank debt in case additional debt is issued under the plan.

  4. Feasible plans would be those that improve the rating of the asset (presently likely to be "C" or "D") such that minimum of the two credit ratings is at or above a threshold level, e.g., at least just below the investment-grade level. The intention is that the asset should not have a high likelihood of ending up in stress soon after restructuring. Therefore, bank debt forgiveness may have to be high enough and its converted equity stake low enough so that new investors can come in with a controlling stake and have incentives to turn the asset around.

  5. Banks can then choose among the feasible plans. Coordination problems can be reduced by employing RBI's Central Repository for Information on Large Credits (CRLIC) and requiring that all plans with 2/3rd approval by outstanding bank credit can proceed. The selected plan would simply be crammed down on any dissenting creditors.

  6. Haircuts taken by banks under a feasible plan would be required by government ruling as being acceptable by the vigilance authorities. Sustainable debt would be upgraded to standard status for all involved banks. The promoters, however, would have NO choice as to what restructuring plan is accepted, and may potentially get replaced and/or diluted, as per the preference of and depending on the price at which the new managing investors come in.

  7. At expiration of the timeline, each exposure that is not resolved will be subject to a steep sector-based haircut for the bank consortium, possibly close to 100%. The promoter will automatically have to leave. These assets would be put into our new Insolvency and Bankruptcy Code regime. Alternately, they could be put up for sale to ARCs and private equity investors who can turn around the assets, levering them up with fresh finance, if necessary. If designed right, only the worst assets should end up in this scenario. However, the possibility of ending up here would serve as a credible off-equilibrium threat so that banks, even the most exposed ones, cannot hold up the restructuring.

There are ways to arrange and concentrate the management of these assets into a single or few private asset management companies (PAMCs), at the outset or right after restructuring plans are approved. These companies would resemble a large private-equity fund run by a team of professional asset managers. Besides bringing in their own capital, they could raise financing from investors against equity stakes in individual assets or in the fund as a whole, i.e., in the portfolio of assets. The portfolio approach might help investors diversify risks on individual assets, improve valuations, and attract greater capital. Bank creditors can set up an oversight committee to ensure cash flows are flowing in and out of the asset restructuring company as per the security rights agreed in the restructuring plans.

Let me emphasize that under this model, the asset management company would be entirely private, like the "Phoenix" structure set up in Spain after 2012 to deal with bank NPAs in Machinery, Steel and Winery.

Let me now turn to

2. Model II: The National Asset Management Company (NAMC). This plan would be necessary for sectors where the problem is not just one of excess capacity but possibly also of economically unviable assets in the short- to medium-term. Take, for example, the Power sector, where projects have been created to deliver aggregate capacity that is beyond the estimated peak utilization anytime soon. Many of these are stalled as they have no fuel inputs and little or no income realization due to lack of credible purchase agreements. Their scrap value is likely small and the only efficient use is as an ongoing concern. If input and output requirements are sorted out, and as power consumption needs rise, these projects could eventually provide value. For a country with per-capita consumption of electricity that is only one-third of the world average, it is reasonable to expect that a well-run power asset won't end up being a white elephant.

Unlike the first model (PAMC) where asset recovery is likely to be relatively quick, these assets may require a long time to start generating cash flows. In addition, the government should have incentives to clear approvals and purchase agreements to make them viable. For both these reasons, such assets would be best quarantined into a national asset management company (NAMC). The NAMC would perform several functions to get the ball rolling: raise debt, say government-guaranteed in part, for its financing needs; possibly raise some more to pay off banks at a haircut, likely steep but softened by payment in the form of security receipts against the asset's cash flows; keep a minority equity stake for the government; and, bring in asset managers such as ARCs and private equity to manage and turn around the assets, individually or as a portfolio. Infrastructure assets that are also long-lived and create externalities (development of townships, improvements in overall productivity, etc.) could be resolved in similar way.

These two models of asset restructuring - one private and the other quasi-government - share many common features with approaches that have been adopted for resolution of stressed assets in history: Sweden ("Securum" and "Retriva") in the early 1990's; in the United States in dealing with the Savings and Loans Crisis ("Resolution Trust Corporation"); in Japan (post-1998 via its Deposit Insurance Corporation); in Indonesia ("IBRA"), Malaysia ("Danaharta") and South Korea ("KAMCO") to deal with the South East Asian crisis; and more recently, in Ireland ("NAMA"), Spain ("Sareb"), and again the United States ("TARP" along with Fannie Mae, Freddie Mac and Federal Housing Administration). In fact, the European Banking Authority has proposed a similar structure to deal with the NPAs of European banks.

Before discussing what all this would imply for bank balance-sheets, let me pose and answer the question: Are these supposed to be "bad banks"? The answer is NO. While I have previously used the phrase "bad bank" for such ideas, over time I have come to dislike the title. A "bad bank" conveys the impression that this entity is to operate as a bank but has bad assets to start with. In fact, the idea is not to operate these entities as banks at all. Resolution agencies set up as banks that originate or guarantee lending have ended up being future reckless lenders, notably in the case of Germany which has often aggregated stressed assets of its Landesbanken into bad banks. I would argue this has also been the story of Fannie Mae and Freddie Mac with respect to housing booms and busts in the United States. It would be better to limit the objective of these asset management companies to orderly resolution of stressed assets with graceful exit thereafter; in other words, no mission creep over time to do anything else such as raise deposits, start a new lending portfolio, or help deliver social programs. It is essential to keep the business model of these entities simple to make them attractive for private investors with expertise for the main task on hand - asset restructuring.

A moment of reflection clarifies that under both models I proposed, bank balance-sheets would be freed up from the overhang of stressed assets and allowed to focus on their healthier activities. The catch is that given the haircuts involved, there will be also be a need for decisive

3. Bank resolution. We keep hearing clarion calls for more and more government funding for recapitalization of our public sector banks. Clearly, more recapitalization with government funds is essential. But few have suggested that the government should adopt measures to economize its total cost. It should ask in return from banks it recapitalizes significant corrective actions, and wherever possible, injections of private capital for loss-sharing with the taxpayers. The expectation of government dole outs might have been set by the past practice of throwing more money after the bad. Take for instance our bank recapitalization plan of 2008-09 after the global financial crisis: banks that experienced the worst outcomes received the most capital in a relative sense to get back to the regulatory capital norms. We must not allocate capital so poorly, recreate "Heads I Win, Tails the Taxpayer Loses" incentives, and sow the seeds of another lending excess.

There are better ways to do it building upon some performance targets already set under the ongoing recapitalization plan. Let me propose five options:

  1. Private capital raising: The healthier public sector banks could have raised private capital by issuing deep discount rights in 2013, and some can still do so now. They must be required to do this to share the government's burden of recapitalizing banks. It might be a good way to restore some market discipline and get their shareholders to more seriously care about bank board and management decisions.

  2. Asset sales: Some banks will have assets or loan portfolios that are in good enough shape to be sold in the market. Assets could be collected across banks and securitized into tranches that are credit-rated, potentially creating some investor demand for buying it at different levels of risk profiles. Such asset sales can generate some of the needed recapitalization.

  3. Mergers: As many have pointed out, it is not clear we need so many public sector banks. The system will be better off if they are consolidated into fewer but healthier banks. After all, we do have cooperative banks and micro-finance institutions to provide community-level banking. So some banks can be merged, as a quid pro quo for timely government capital injection into the combined entity. It would offer the opportunity to rejig management responsibility away from those who have under-performed or dragged their feet the most. Synergies in lending activity and branch locations could be identified to economize on intermediation costs, allowing sales of real estate where branches are redundant. Voluntary retirement schemes (VRS) can be offered to manage headcount and usher in a younger, digitally-savvy talent pool into these banks.

  4. Tough prompt corrective action: Undercapitalized banks could be shown some tough love and be subjected to corrective action. Such action should entail no further growth in deposit base and lending. This will also restore some market discipline in deposit migration, away from the weakest public sector banks that have price to book equity (P/B) ratios of around 0.5 or lower, to healthier public sector banks (P/B ratios around 1) and private sector banks (P/B ratios of 1.5 to 4.5). The market has given its verdict as to where the growth potential in our banking sector lies and deposit growth should be allowed to reflect that.

  5. Divestments: Undertaking these measures would improve overall banking sector health and market-to-book valuations, creating an opportune time for the government to divest some of its ownership of the restructured banks. This would also reduce the overall amount the government needs to inject.

There are many details to work out. But I hope this provides a start. It is going to require being balanced and creative, holistic and uncompromising, in achieving the end goal. Piece by piece approach with all discretion given to banks simply hasn't worked. Time is of the essence if we are to restore corporate investment and job creation.

Sustainable progress in an economy cannot occur when a set of players is allowed to hold up the efficient allocation of capital. Their owning a smaller share of assets can help unlock economic value; their hogging of these assets will only lead to further value-erosion.

I hope we can work together and collectively make the right choice.

Thank you for your kind attention.