Friday, February 5, 2016

Bad loans and bad banks

The Government of India is apparently contemplating the establishment of a 'bad bank' to isolate the distressed assets held by the banking sector. Essentially bad banks are asset management companies which purchase distressed loans at knock-down prices and then revive and sell them to investors.

I had blogged about bad banks and other bank resolution options here and here. At a fundamental level, a bad bank is best captured by the graphic below. Incidentally, this classic model of resolving distressed assets of private banks failed to find traction even in the US during the height of the sub-prime crisis for lack of consensus on what would be an acceptable price discovery mechanism.
There are atleast four models of resolving bad assets. At one end is the 'Swedish model' of complete nationalization, whereby the equity holders are stripped-off their holdings and the assets transferred to one public aggregator "bad bank". This bank can then restructure and manage these assets back to health and dispose them. Alternatively, as Citibank did, each existing bank can create its own "bad bank" and then manage those assets. Another option, commonplace in the US, is to sell/auction off such assets to Asset Reconstruction Companies (ARCs) or even fund managers, with or without partial government guarantees. The least disruptive option is to simply securitize such loans and sell them off, with some sovereign guarantee as is being proposed in case of Italy. A snapshot of all these options is below.
During the Eurozone crisis, bad banks were established across many countries. The three largest - Ireland's NAMA, Sapin's Sareb, and Germany's FMS Wertmanagement - were established as publicly owned entities with 10-15 year life-span, who purchased, largely real estate loans, at haircuts, with the objective of reviving and selling the impaired assets and recovering some value. Their experience has been mixed.

In India, the debate has been whether the distressed assets should be resolved through a publicly owned 'bad bank' or through sales to ARCs. In this context, a few observations.

1. In India's case, the vast majority of assets are held by public sector banks. So any haircuts by the bank are borne by the government itself. This raises the question of whether an aggregator 'bad bank' is the most efficient and incentive compatible way to assume losses.

2. In the US and western Europe, where the vast majority of assets were securitization products, mortgages, real estate, consumer, and student loans. Being financial or functioning assets, at worst, these assets could simply be written off. In contrast, Indian banks' problems stem from commercial loans to real estate and infrastructure project developers, a large proportion of which are non-income generating and at different stages of completion. Here, the completion of these projects is as much important as the bad loans resolution itself. But this would, in most cases, require further capital infusion. In other words, it is important to keep in mind that the resolution of such loans through a bad bank is not a substitute for restructuring distressed projects, as much important for the government as repairing bank balance sheets.

3. The credibility of the process assumes great significance. The markets have to be convinced that the resolution process has been done rigorously enough so that the "good" bank is a healthy enough institution. This would require that all the distressed assets are hived off and those done so are valued fairly.

4. On this, being public institutions, the resolution process runs the risk of being plagued by risk-aversion. The administrators of the process may unwittingly structure the process to maximize the returns for the banks. This corollary effect of this would be increased riskiness of the hived-off asset as well the lower likelihood of its successful resolution. A public sector decision-making process may not be best suited for managing a price discovery process that equates the interests of all sides.

5. Apart from risk-aversion, the public sector banks are certain to be hobbled by the decision paralysis syndrome that arises from the risk of post-facto vigilance and audit oversight. In a regime constrained by audit and other oversights, it may not be possible to sell off all assets. It may be necessary to exercise judgment and write-off some assets, or even sell them off at prices far lower than any internal 'off-set' price.  

6. Once decided, this must be done very quickly. Else the possibility of generating moral hazard among banks and other market participants is considerable. This is all the more important given the well-known fact that a not insignificant share of the restructured loans under CDR and 5:25 scheme are not exactly performing assets.

7. But this is most likely to run into the problem of market demand for such assets, given the flood of such assets likely to be released into the market.

8. Finally, it may be necessary to examine the reasons why private ARCs have not made much headway. There may be a need for regulatory enablers like uniform provisioning norms for various kinds of asset restructuring mechanisms.

Given all this, what is the way out for India. The fundamental question that we should be asking is - What is the most effective strategy to resolve the bad loans in bank balance sheets while simultaneously completing the stalled projects which form a major share of these non-performing assets? This requires going much beyond blindly replicating 'bad bank' models elsewhere.

For a start, it would be ncessary to segregate and classify the different categories of loans. Would the under construction infrastructure assets be best resolved by remaining in the respective bank balance sheets and restructured as individual projects, with promoters and lenders taking haircuts, and government infusing capital? Or would they be best resolved by transferring all of them into one entity and then restructuring them? Which of the two approaches would best align incentives and minimize transaction costs, given the constraints of public sector decision-making in India? While there are pros and cons with both approaches, I am inclined to the former.

Given that the broader real estate and infrastructure sector loans form the major share of distressed assets, it may be reasonable to confine this resolution process to them. The other smaller loans can be dealt separately, either written off fully or sold to ARCs. This sectoral focus would enable the asset managers to more effectively manage and resolve them. Further, their  "fair" valuation becomes far less of a challenge with such assets.

All this does again raise the question of recapitalization. Bad assets resolved or not, Indian banks need massive recapitalization, far higher than the drips being administered now. It may be even worthwhile to dedicate a full 0.5% of GDP fiscal deficit relaxation this coming budget exclusively for recapitalization complemented with a focused distressed assets resolution mechanism. 

See these reports from IMF, Bruegel, and McKinsey for more on resolution of distressed assets and bad banks. 

No comments: