The previous “Noise to Signal” column titled The Silent Role of Credit Ratings in India’s Bad Loan Crisis highlighted how independent, private sector credit rating agencies and equity analysts issued glowing, positive ratings to all the corporates that indulged in heavy borrowing during the years 2010-2014.
It argued that in this backdrop, it is perhaps inaccurate to claim that India’s bad loans crisis is entirely due to a nexus between public sector banks (PSU) and crony businessmen. The column received wide feedback, with some wondering how private sector banks seem to have emerged unscathed in the current bad loans crisis.
As of financial year 2014-15, reported non-performing assets (NPA), aka bad loans, of private sector banks were roughly 1.6 percent of their total assets while that of PSU banks were nearly three times worse at 4.5 percent. India’s private sector banks, on average, command a valuation of nearly 2 times their book value in the stock markets. On the other hand, government owned PSU banks are valued significantly below their book value. This discrepancy in NPA levels is the justification for the large valuation gap between private sector and PSU banks. Thus the argument, that while credit rating agencies, equity analysts and PSU bankers may have erred in judgement, private banks seem to have been more prudent in their lending decisions.
Comparing Small Oranges and Watermelons
Except, comparing NPAs of private sector banks with PSU banks’ is like comparing small oranges and watermelons. This comparison ignores the fact that the private sector banks hardly lend to corporates vis-à-vis PSU banks. It is then natural that NPAs of private sector banks will be much lower than those of PSU banks, since they have largely shunned risky corporate lending. To be clear, this is not a defence against the alleged malpractices in PSU bank lending. This is an attempt to infuse some data backed context to the narratives around India’s bad loans crisis.
Private Sector Banks Play It Safe?
Six public sector banks (State Bank of India, Punjab National Bank, Bank of India, Union Bank, Indian Overseas Bank and Bank of Baroda) and seven private sector banks (ICICI Bank, HDFC Bank, Axis Bank, YES Bank, IndusInd Bank, Federal Bank and Kotak Mahindra Bank) combined, account for nearly two-thirds of all of India’s banking. Indian corporates borrowed roughly $2 trillion in loans from these 13 banks in the financial years FY10 to FY14. Nearly $1.5 trillion of these came from the six PSU banks and only the remaining $500 billion from the seven private sector banks. 75 percent of India’s corporate loan requirements were met by PSU banks while the private sector banks accounted for a mere 25 percent.
In the same period, FY10 to FY14, Indian consumers borrowed roughly $650 billion from these 13 banks. Of this $325 billion came from the seven private sector banks and the remaining $325 billion from the public sector banks. That is, Indian consumers are financed equally by private sector and PSU banks. While corporate loans are catered to largely by PSU banks, consumer loans are split equally between PSU and private sector banks.
Loans to businesses are ostensibly far riskier than loans to consumers. Consumer loans are typically well secured through regular income streams of individuals and are backed by assets such as homes or jewellery. Recovery of corporate loans, on the other hand, is dependent entirely on the success of the business plans for which the money has been borrowed. In general, NPA levels in consumer loans in India have tended to be less than half of NPAs in corporate loans, over the past few decades. In other words, there is twice the chance that a loan to a corporate will turn sour vis-à-vis a loan to a consumer.
The chart below shows the relationship between each bank’s propensity for corporate loans relative to its size and its gross NPA levels, for the banks in our study. The size of the bubble represents each bank’s share of overall corporate loans in the banking sector. Notice how as a bank gives out more in corporate loans vis-à-vis consumer loans, its NPAs also rise. It is thus prudent for any lender to maximise its share of consumer loans and minimisecorporate loans in its portfolio to mitigate risk of bad loans and NPAs.
It is also true that the size of private sector banks is much smaller compared to PSU banks. Even if one adjusts these banks for their relative sizes, it is very clear that the private sector banks shun corporate lending and prefer to lend to consumers. The ratio of corporate loans to consumer loans for the seven private sector banks in this time period is roughly 1.5, i.e. private banks lent Rs 150 to corporates for every Rs 100 to the consumer. While PSU banks lent Rs 700 to corporates for every Rs 100 to consumers. So even relative to their sizes, it is clear that the private sector banks prefer to lend to consumers and not to corporates to the extent that PSU banks do.
The Path to Higher Valuation
The private sector bank, Kotak Mahindra Bank, reported FY15 gross NPA of 1.9 percent while PSU, Union Bank reported a 5 percent gross NPA, 2.5 times worse. In the previous five years from FY10 to FY14, nearly 8 percent of all corporate loans were given by Union Bank while only 0.7 percent were given by Kotak Mahindra Bank.
Kotak Mahindra Bank only lent Rs 120 from its books to corporates for every Rs 100 lent to consumers. On the other hand, Union Bank lent Rs 800 from its books to corporates for every Rs 100 to consumers. The stock market values Union Bank at half its book value while Kotak Mahindra Bank is valued at nearly 4 times its book value, i.e. Kotak Mahindra Bank is valued EIGHT times higher than Union Bank. Clearly, the path to higher valuation seems to be to shun riskier corporate lending as much as possible.
However, if one were to ask the question – had Kotak Mahindra Bank indulged in similar share of corporate lending, as say Union Bank, relative to its size of assets, would its NPAs still be as low? Of course, it’s a purely hypothetical and an academic question but nevertheless, one that can perhaps be insightful.
‘Normalising’ the NPA Data
Kotak Mahindra Bank’s ‘normalised*’ NPA is around 3.4 percent, while Union Bank’s is 5 percent, worse but not 2.5 times worse, as the reported NPA numbers indicate. In other words, had Kotak Mahindra Bank indulged in a similar share of corporate loans vis-à-vis consumer loans as Union Bank, its NPAs could have been 3.4 percent and not the reported 2 percent.
Similarly, ICICI Bank’s reported FY15 NPA of 3.8 percent could have jumped to 6.4 percent, (same as Punjab National Bank’s FY15 reported NPA) if its ratio of corporate to consumer loans mirrored those of PSU banks.
True that this is a simplistic extrapolation exercise and does not imply that these private sector bank NPAs would have actually been at these levels for similar share of corporate lending. But the point is that a comparison of NPAs between private sector and PSU banks is too simplistic.
Normalising for both a bank’s share of overall corporate loans in the system as well as a bank’s own share of corporate loans relative to its book, the charts below show that while PSU banks still have larger NPAs than private banks, the difference in NPA levels between PSU banks and private sector banks is not as large as their stock market valuations seem to imply.
Notice in the charts, how on a normalised basis, four PSU banks perform better than the private sector bank – ICICI Bank. But the stock market values ICICI Bank at 2.5 times more than these four PSU banks. It is true that this analysis uses FY15 reported gross NPAs and the stock market could have well factored in future deterioration of these NPA numbers for PSU banks.
Mirror, Mirror on the Wall
HDFC Bank is one of the world’s most valued banks, trading at nearly four times its book value. It is largely because HDFC Bank religiously shuns corporate lending. Despite being one of the largest private banks in India, it contributed less than 5 percent of overall corporate loans since FY10 while it accounted for 20 percent of overall consumer loans. In contrast, the largest public sector bank, State Bank of India, accounted for 35 percent of all corporate loans in this period and an equal share (20 percent) in consumer loans.
Admittedly, this normalised NPA analysis does not factor in behavioural changes in lending practices and the possibility that private sector banks get to cherry pick corporates given their low share of corporate lending. Also, those that are turned away by private sector banks tend to borrow from PSU banks, leading to adverse selection impact on PSU banks. Then, to improve efficiency, should India’s public sector banks also eschew corporate lending and stick to consumer lending, largely?
If say, State Bank of India had adopted HDFC Bank’s ratio of corporate to consumer loans during the period FY10 to FY15, it would have meant $500 billion less capital for Indian corporates in this period. Perhaps SBI’s NPAs today would have been much better. But with no other financing alternatives, such as robust corporate bond markets or other debt investors for corporates to borrow this $500 billion from, one wonders if India could have still built the airports, roads and power plants that it did in those years.
Sure, some of these loans went awry. Sure, some may have even been crony lending. Surely, some of these riskier corporate loans could have priced in the risk adequately. However, it is quite likely that the very risky nature of corporate lending vis-à-vis consumer lending combined with the fact that only the PSU banks indulge in such heavy corporate lending explain the higher absolute NPA levels of PSU banks versus the private sector banks.
This is certainly not to condone perpetual recapitalisation of PSU banks with taxpayer money for their bad loans sins. But to simplistically argue that PSU banks should merely imitate private banks in their lending behaviourfor efficiency and valuation reasons may perhaps be akin to cutting one’s feet to fit into small shoes rather than to buy new, larger shoes, as the Chinese saying goes.
Published on 19 August, 2016 in Bloomberg Quint