Wednesday, November 21, 2018

Credit stress in large Indian firms

by Ajay Shah and Pramod Sinha.

We in India are used to thinking about banks and NPAs. We infer the state of difficulty of the banks, and indirectly of their borrowers, by using data from banks. There are many advantages in looking directly at the state of credit stress in the large non-financial firms, and identifying the firms where there is high credit stress:

  1. Do not rely on bank information. This evidence is not filtered through the difficulties of banking regulation. Whether a bank classified Kingfisher Airlines as an NPA or not, we can see credit stress in the financial statements of Kingfisher Airlines.
  2. Look beyond banks. Banks are not the only financial lenders to the non-financial firms. As an example, bank-centric thinking is not useful in understanding runs on mutual funds. When there is stress in a borrower, this impacts not just on banks but on all lenders. Pulling together information about stressed borrowers helps us see the difficulties of lenders, on a financial system scale, and not just in banks.
  3. Micro-prudential considerations. The stressed firms face likely defaults. The debt of such firms is likely to be worth less than book value. Under sound micro-prudential regulation, banks and other lenders should mark down these assets, even if no default has taken place. The extent of stress, as seen here, gives us insights into the fragility of banks and other lenders.
  4. The bankruptcy process, the distressed debt industry. There is a new world opening up in India, of distressed firm transactions and the bankruptcy process. We will see the empirical contours of this industry, and the bankruptcy process, by examining the state of credit stress in the non-financial firms.
  5. A drag on growth. A firm that is in a state of credit stress is likely to face difficulties meeting payments to creditors. It might often be liquidity constrained, and may struggle to obtain cash to pay its suppliers. The mind space of the leadership of such a firm is likely to be absorbed in the struggle for survival. Such firms are unlikely to fare well on growth through increasing the resources utilised or through increased productivity. To understand what is coming in Indian macroeconomics, we should look at the non-financial firms and their balance sheet difficulties.

Identifying stressed firms

The interest cover ratio is defined as PBIT/interest. If a firm has to make interest payments of 100, and if its profit before interest and taxes is 150, then its interest cover ratio ("ICR") is 1.5. Such a firm has the 100 required to pay interest in the year, but there may be a task in terms of juggling the dates on which interest has to be paid versus the dates on which the business produces cash. And, such a firm is left with just 50 after paying interest, which can be used for debt repayment and the regular capital expenditures required for the upkeep of the business.

A good thumb rule which identifies a firm in a state of stress at time $t$ is: The firm has ICR$ < 1.5$ in year $t$ and in year $t-1$. This avoids the false positive of a firm which only hits ICR$<1.5$ for one year.

A `stressed firm', by this definition, is not necessarily one that has defaulted (and is thus eligible for the bankruptcy code), and it is not necessarily one that is classified as a non-performing asset by RBI's rules of recognition. We would, however, suggest that a firm with two consecutive years at an ICR of below 1.5 is under stress, has an enlarged risk of default, and has a management team that is absorbed in dealing with this stress.


We study all the non-financial firms in the CMIE database. At each year, we isolate the firms which are observed for two consecutive years. Some additional sanity checks are applied. Through this, we are able to construct two sets at each point in time: The set of all firms observed and the subset of this, which is the stressed firms.

Here are some counts of the firms in the two sets.

2014-15 9,289 3,674
2015-16 9,208 3,702
2016-17 6,687 2,573

In the table above, the total number of firms (9,289) for 2014-15 is the number
of non-financial firms that are observed, and pass some sanity checks, in both 2013-14 and 2014-15. Of these, 3,674 were stressed. The last year that we utilise here -- 2016-17 -- has fewer firms when compared with the years prior to
it, where information for a larger number of firms has trickled into the CMIE database. In this last year, we see 2,573 non-financial firms in the database, where the ICR was worse than 1.5 in both 2015-16 and 2016-17.

Conditions in 2016-17

Parameter Value (Rs. Tln)
Balance sheet size
   Stressed firms 29.79
   All firms 77.24
Bank borrowing
   Stressed firms 8.87
   All firms 14.94
Total borrowing
   Stressed firms 15.58
   All firms 27.59

This shows that the sum of the balance sheet size for all the 6,687 firms for 2016-17 was Rs.77.24 trillion. Of this, Rs.29.79 trillion was in the 2,573 stressed firms.

Totally, borrowing of Rs.27.59 trillion was visible. This is small when compared with the total assets of these firms of Rs.77.24 trillion. Of this borrowing, Rs.15.58 trillion was in the stressed firms.

Finally, we are able to see Rs.14.94 trillion of borrowing from banks, in this sample of 6,687 firms, in 2016-17. Of this, Rs.8.87 trillion was in the stressed firms.

How has credit stress evolved over time?

We are able to do these calculations for all years from 1998-99 onwards. We will express the time-series evidence using confusingly similar graphs, all of which produce important stylised facts for our understanding of the economy.

The share of bank debt to stressed firms, in the total bank debt seen in the sample firms

The health of banks is related to the health of their borrowers. Hence, in the graph above, we compare the sum of bank credit to stressed firms (in our sample) against the sum of bank credit to all firms (in our sample).

The business cycle is clearly visible here. In the last tough downturn, 2000-2003, this ratio was at about 50%. That is, about half of the bank borrowing seen in the CMIE database was in stressed firms.

This ratio dropped all the way to about 15% in 2007. It climbed steadily thereafter and is now at about 60%. There is a tiny gain in 2016-17 when compared with the previous year.

In the last recession, this measure improved through the recovery of the economy. Firm exit took place through the sluggish traditional ways. When the bankruptcy reform resolves or liquidates a large volume of stressed firms, this will deliver improvements in this measure. To the extent that the bankruptcy reform works, we may expect the next recovery to proceed faster than the last one, where it took five years of a powerful expansion to get from about 50% to about 15%.

We apply this same thinking to total borrowing -- instead of just bank borrowing:

The share of total borrowing by stressed firms, in the sum of borrowing seen by all sample firms

The last business cycle downturn got to values of above 50%, there was a great decline to about 20%, and then it has risen to about 55%, with a slight improvement in 2016-17.


There is considerable balance sheet stress. In the latest year, the aggregate balance sheet size of the stressed firms -- observed in the CMIE database -- was Rs.30 trillion. The stressed firms had Rs.15.6 trillion in borrowings of which Rs.9 trillion was from banks. This has implications in other parts of finance, beyond banking.

Bank debt in stressed firms is about 60% of total bank debt seen in the sample. Similarly, the borrowing by stressed firms is about 55% of all borrowing in the sample. Under sound micro-prudential regulation, asset-based lenders would mark down these assets based on the price at which the loan/bond could be sold on the market.

In the conventional wisdom, there is about Rs.10 trillion of bad debt on the balance sheet of banks. Our analysis shows that in the 6,687 large non-financial firms, where Rs.15 trillion of bank debt is located, we see 2,573 stressed firms with Rs.9 trillion of bank debt. The 2,573 stressed firms that we see in this sample, alone, account for 11.4% of the overall bank debt ("non-food credit") in the economy.

The stressed firms are about 40% of the overall corporate balance sheet. These firms are likely to fare poorly in investing or in productivity growth, and are thus a drag upon overall economic growth.

It is likely that many of these stressed firms will be sold, or go into the bankruptcy process. There is a substantial task ahead, in terms of resolving these firms and paying for the losses experienced. These 2,573 firms are the happy hunting ground for this new industry. This process of resolution is central to India's economic recovery.

There is much value in understanding the balance sheet stress in the economy using such methods. We obtain insights into difficulties of the financial system going beyond a bank-centric view, we get a view of the new distressed debt industry, and we get insights into the drag on GDP growth that the stressed firms represent.

The authors are researchers at NIPFP.


  1. The author does not provide any empirical evidence from the database about the accuracy of the the mentioned thumb rule (<1.5 implied stress).

    If the rule is correct, proportion of defaulters within stress segment (<1.5) should be higher than the proportion of defaulters within non stressed segment (>1.5). In other words, any firm with pbit/interes less than 1.5 is more likely to default than a firm with pbit/interest more than 1.5.

    In absence of any such backtesting of the thumb rule, the mentioned numbers for stressed segment (as a hunting ground for bankruptcy) seems too large to believe.

  2. why to think of single variable thumb rule to identify future defaults.

    The general and prevalent thing would be to simply develop a probability of default model using multiple accounting ratios reflecting not only coverage but also dimensions such as activity, leverage, liquidity, profitability etc.

  3. We are not discussing defaults. We are discussing credit stress.

    There is no publicly accessible defaults database in India so it is not yet possible to estimate credit default models. In any case, that's not the central point here in this article.

    ICR<1.5 for 2 years running is a statement of substantial credit stress. This is well understood in the global literature on credit risk. It is also commonsense - think of the troubles of the firm dealing with life with only Rs.50 of spare capacity.


Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticism is perfectly okay; uncivilised language is not. We delete any comment which is spam, has personal attacks against anyone, or uses foul language. We delete any comment which does not contribute to the intellectual discussion about the blog article in question.

LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.