Money & Finance

What ails the Indian banking sector?

  • Blog Post Date 20 September, 2018
  • Perspectives
  • Print Page
Author Image

Pronab Sen

Chair, Standing Committee on Statistics

Non-performing assets (NPAs) of Indian banks have risen from under 3% to over 13% in the past couple of years, making the state of the banking sector one of the biggest challenges facing the country in accelerating investments and growth. In this post, Dr Pronab Sen diagnoses the causes of the current NPA problem, and proposes solutions.


The past two years or so have witnessed rapidly increasing stress in the Indian banking sector, with non-performing assets (NPAs) steadily climbing from under 3% to over 13% of total assets. Loan-loss provisioning for these NPAs has seriously eroded the capital base of a number of banks, thereby limiting their ability to make further loans.1 There is now general consensus that the state of Indian banking is one of the biggest challenges facing the country in accelerating investments and growth.2 Expectedly, given the magnitude and gravity of the problem, much has been written and debated about ways of tackling current NPAs and of preventing recurrence of such a problem in the future.3

There is certainly urgency in finding a solution to the current NPA problem, but diagnosing the cause(s) and putting correctives in place is perhaps even more important. Unfortunately, the entire discourse around this issue appears to have boiled down to the governance practices in public sector banks (PSBs), particularly political interference and crony capitalism. The solutions, therefore, have more or less polarised around two ideologically coloured views – (a) privatisation of PSBs; or (b) strengthening independence and capacity of PSBs. While this is a perfectly legitimate debate and may well lead to a part of the solution, it should not forestall the search for other causes, which may be just as important.

The cause

The purpose of this article is to argue that much of the problems currently being faced by the banking sector arise from a fundamental change that occurred in the Indian banking sector after 2002. Prior to that, Indian banks mainly had two types of loans: (a) working capital loans to production entities, firms, and farmers, which accounted for 76% of the banking portfolio; and (b) retail term loans to households for housing, and durable goods, which was somewhat less than 24%. Since then banks have been aggressively making term loans to companies for fixed capital investments, such as land, building, and machinery. This now accounts for 38% of the banking portfolio, with working capital at 42% and retail at about 20%. It so happens that the bulk of the NPAs by value are long-term loans to corporates.

This shift from short-term working capital loans to long-term loans has not gone entirely unnoticed. Concern has been expressed in many quarters about the serious asset-liability mismatch in the banking sector, whereby the average tenor of assets (read loan portfolio) has been rapidly going up relative to that of the liabilities (deposits). While this is certainly a matter of concern, it is not the cause of the NPA problem. 

What is missed by the commentators is that there is a significant difference between the de jure (contractual) and the de facto (behavioural) tenors of bank deposits. The average de jure tenor is usually computed by taking the weighted average of the contractual tenors of demand and term deposits, which works out to around two years.4  However, most deposits, both demand and term, usually remain with banks for much longer periods, which can be as high as 14 or 15 years. Thus, the de facto average tenor is probably above 10 years, which may well be higher than the average maturity of the loan portfolio. This difference between the de jure and the de facto means that the asset-liability mismatch becomes a problem only when there is a loss of depositor confidence and withdrawal of deposits.5 This is not the situation as yet.

In my opinion, the NPA problem has arisen from four features that characterised bank loans for fixed capital formation:

  • Banks did not, and even today do not, have the capacity to assess the long-term credit-worthiness of borrowers. They rely almost exclusively on assessments made by credit-rating agencies. Unfortunately, the credit-rating agencies provide ratings only when the company concerned intends to raise debt capital directly from the market. Therefore, in cases where the company has not issued long-term bonds, there is no real basis to judge the viability of the company.6
  • Banks have no capacity to monitor the use of the long-term loans by the borrower. This is particularly egregious for project loans, especially when it is without recourse.7 The consequence of this inability is that the parent company can use the funds for purposes other than for which they were borrowed without the banks being any the wiser. As a result, whatever little appraisal was done initially becomes seriously vitiated.
  • Most importantly, such loans create an existential problem for the borrower since attaching such assets essentially means “winding up” the companies.   It is only natural that company promoters would fight such an eventuality tooth and nail. Since winding up petitions, especially with multiple lenders, was a complex legal process, which frequently took 7-10 years in the courts, with promoters using every possible trick to stall proceedings, banks faced the alarming prospect of making large loan loss provisions which would stay on the books for many years with very little recovery at the end. It is no wonder that bank managements have allowed these potential NPAs to simply accumulate over the years through myriad forms of “ever-greening”.8 Reserve Bank of India’s (RBI) forced recognition of these NPAs has now brought the problem to light.9
  • Finally, there was gross mispricing of loans by the banks. In practically all banks, the yield curve is almost parallel to, and sometimes even flatter than, the yield curve on government securities. Even if the banks were unable to assess the appropriate risk premium at different tenors due to (a) above, there is no reason why a premium was not attached to the illiquid nature of the underlying collateral, which is also related to (c) above. 

The origins

This high exposure to companies in fixed assets, especially in project finance, did not happen by itself, but at the behest of successive governments. It really begins with a decision taken by the government in 1997 to stop issuance of tax-free bonds by the Development Finance Institutions (DFIs)10, which were the main source of term finance for corporates. This closed the only source of relatively low-cost funds for these institutions, and was instrumental in some of them converting to commercial banks and others shutting down. Consequently, Indian commercial banks were forced to fill the vacuum and, in effect, convert themselves to “universal banks”.11 

The problem escalated due to the government’s focus on infrastructure during the 2002 to 2009 period, especially with the efforts made in promoting ‘public-private partnerships’.12 This led to a rapid increase in project financing to private infrastructure firms. The global financial crisis in 2008-09 again required the banks to take substantial further exposures in long-term loans as external loans taken by Indian corporates simply dried up.13 Therefore, to pin all the blame on poor governance in PSBs, especially by government interlocutors, appears gratuitous. 

The solutions

The first and, by far, the most important solution to the NPA problem is to provide for rapid resolution of defaults. It should be remembered that something similar used to happen earlier with working capital loans as well. This was taken care of by the promulgation of the Securitisation and Reconstruction of Financial Assets and Enforcement of Security Interest (SARFAESI) Act in 2002, which resulted in gross NPAs being reduced from 19% in 2002 to 6% by 2006. However, the SARFAESI Act is of consequence only in cases where the collaterals are moveable assets, and especially where there is a single claim-holder. It is simply not appropriate for fixed capital loans.

The need for a Bankruptcy Act was anticipated and recommended in the Tenth Five Year Plan as far back as 2002. After much delay, the Insolvency and Bankruptcy Code (IBC) was eventually passed in 2016, which has now made it much easier for banks to effect recovery through liquidation of fixed assets. As the IBC institutions and processes are strengthened, it should take care of the existing NPAs and, up to a point, the need for banks to do ever-greening. However, it does not in itself address the other issues associated with long-term loans, especially that of mispricing.

The other critical solution is fundamental to the change that has taken place in Indian banking, and which is not even under discussion. Indian banking laws follow the “Anglo-Saxon” model in which there is a firewall between commercial and investment banks. The shift to universal banking should have been accompanied by amendments to these laws, but we chose to follow a ‘reform by stealth’ approach.14 The key change is that countries which have a universal banking model, such as Europe and Japan, permit banks to issue bonds for financing term loans. Our Banking (Regulation) Act 1949 does not do so. Banks can issue bonds only for raising their own capital. 

The only way to correct this problem would be to amend the Banking Act to permit issuance of bonds by banks for on-lending.15 Ideally, banks should be required to ensure that a minimum percentage of their term loans are financed by long-term market borrowings. Such a measure would have a number of positive effects. First, if the banks are required to issue bonds to finance a part of their term loans, it would lead to a much more rational (read steeper) yield curve on bank loans since the interest rate on bank bonds would be significantly higher than those on deposits.

Second, a more rational yield curve would also probably drive some corporates to borrow directly in the bond market, thereby reducing the pressure on banks. The Security and Exchanges Board of India (SEBI) has made it mandatory for listed companies to raise a part of their long-term debt from the market, but this can potentially be very distortionary in terms of credit allocation in the economy.16 A more rational bank yield curve would obviate some of this distortion since there would be a self-selection element in companies’ decision to issue long-term bonds.17

Third, this would automatically correct the growing asset-liability mismatch problem, and thereby improve the credibility of the banking sector as a whole.

Finally and perhaps most importantly, since banks would have to get themselves periodically rated, the yield curve of individual banks would reflect the strength of their loan portfolios. As a result, banks with weaker balance sheets will have to offer higher interest rates on their bonds, which will get reflected in higher interest rates that they will have to charge to the same potential client. This should have the effect of introducing a self-correcting mechanism in the lending behaviour of banks well before they run into serious problems.

Thus, the choice facing the government is clear: either revert to the status quo ante where banks eschew lending for fixed capital formation;18 or amend the banking laws to reflect the new reality. Sitting on ones hands and berating the banks is not an option.


  1. Banking laws require banks to limit their loan portfolios to a particular multiple of their capital base, referred to as the Capital Adequacy Ratio (CAR).
  2. The other two major challenges being: (a) the precarious state of the world economy; and (b) rural distress.
  3. A simple web search reveals that there are more than 100,000 articles on how to tackle the NPA problem in Indian banks and more than 200,000 on its causes.
  4. Demand deposits (termed CASA (current and savings account) in banking parlance) can be withdrawn immediately, that is, they have a contractual tenor of zero, and term deposits have contractual tenor that varies between one and five years. Since CASA usually is between 35 to 40% of total deposits, the average comes to around two years.
  5. A ‘bank run’ is an extreme case of this.
  6. Most corporates do have ratings for shorter durations, and banks tend to use these as proxies for long-run risk.
  7. ‘Without recourse’ means that the lender has no claim on the assets of the parent company, but only on those of the project (which are usually developed through special purpose vehicles (SPVs)). Most infrastructure loans fall into this category.
  8. Since balance sheets and profit and loss accounts are the basis on which management is evaluated, ‘kicking the can down the road’ is a time-honoured practice of top management in non-promoter driven companies, which almost all banks are.
  9. The ‘poor governance in PSBs’ narrative, which hinges around higher NPAs in PSBs relative to private banks is driven by two, not necessarily independent, factors: (a) private banks tend to have lower proportion of project loans than PSBs; and (b) some private banks are in fact promoter-driven.
  10. The main DFIs were Industrial Credit and Investment Corporation of India (ICICI), Industrial Development Bank of India (IDBI), Industrial Finance Corporation of India (IFCI), and Industrial Investment Bank of India (IIBI).
  11. Universal banks perform the roles of both commercial banks and investment banks.
  12. At this time, infrastructure was included in the ‘priority sector’ list.
  13. There had been a sharp increase in external commercial borrowings by Indian corporates, which rose from about US$3 billion in 2002 to US$28 billion in 2008.
  14. The US, which also followed the same model, adopted a legislative approach to move to universal banking by diluting the Glass-Steagal Act. The UK too acted similarly. In both cases, there has been a reversal after the global financial crisis, which we would do well to bear in mind.
  15. It surprises me that RBI has not demanded this change. I can only ascribe it to ‘lazy’ regulating – to use RBI’s own expression.
  16. The logic of the move, at one level, is sound in that it will force all listed companies to obtain long-term ratings, which will be useful not only for banks but also for equity investors. However, it can work against younger companies, especially those which are recently listed.
  17. This could of course lead to an adverse selection problem, which implies that banks need to improve their appraisal capabilities in any case.
  18. In all fairness, over the years the government has tried to create new institutions for infrastructure lending, but with a singular lack of success.



By: icicibank customercare

your article is very good and full of information thanks for sharing

By: Ramakrishnan Sundaram

Even the Sarfaesi Act 2002 was not very efficient in resolving working capital NPAs, with recalcitrant borrowers gaming the system. At the end of the day, all these problems were eventually caused by the flooding of Indian bourses with dollar funds during QE which drove down the repo rate and their reverse flight during taper tantrums leading to higher repo rate. Apart from the issue of Bank bonds to finance long gestation projects, we should disincentivise such hot money from abroad.

Show more comments
Join the conversation
Captcha Captcha Reload

Comments will be held for moderation. Your contact information will not be made public.

Related content

Sign up to our newsletter