It is tempting to believe, like RBI, that the worst of the NPA crisis is behind us and that banks, especially PSU ones, are on the road to recovery. In its latest Financial Stability Report (FSR), RBI says that NPAs (as a percentage of bank loans), which rose from around 7.5% in March 2016 to as much as 11.5% in March 2018, fell to 10.8% in September 2018, and are likely to fall further to 10.3% by March 2019 and then to 10.2% in September 2019.
In a severe stress scenario, gross NPAs are projected to rise to 10.8% in September 2019, but how remote this possibility is, according to RBI, can be seen from the assumptions made. A severe stress situation is one where, for instance, GDP growth collapses to 5% in FY19 and then to 4.4% in FY20 while the fiscal deficit worsens to 5% (as compared to the budget target of 3.3%) and then to 5.4% in the same period.
Since it is unlikely India is going to be in such a high-stress situation, the overall message is one of comfort. While that is comforting, it is best to be cautious since, for one, RBI has got it terribly wrong in the past. In the June 2014 FSR, for instance, RBI’s NPA projections for March 2015 were 4% as the baseline and 5% in a severe stress scenario! Mind you this is when, for several years, Credit Suisse’s House of Debt series was talking of highly stressed corporate balance sheets; the August 2012 report, for instance, talked of how 10 stressed borrowers accounted for 13% of banking sector loans and their combined interest cover—ebit-to-debt—was just 1.5.
Even in the December 2015 FSR, RBI was projecting a 4.9% NPA level for March 2016 and a 5.2% level for September 2017 in the baseline scenario. The actual NPA levels, it turns out, was 7.6% in March 2016 and 10.2% in September 2017. That RBI should have got it so wrong (see graphic) is not surprising since the Credit Suisse report of October 2015 was talking of problem loans being as high as 16.6% of total loans even as the gross NPAs—as shown by RBI data—was a mere 4.5%.
It was only in June 2016 that the FSR started projecting somewhat more realistic numbers after the Asset Quality Review (AQR); as a result of the AQR, gross NPAs rose sharply to 7.6% of gross advances in March 2016, from 5.1% in September 2015. Even after this—stressed assets were seen at 11.5% of advances—RBI projected NPAs rising to just 8.5% in March 2017, and rising to 9.3% in a severe stress situation.
And, in December 2017, RBI projected September 2018 gross NPAs at 11.1%, going up to 12.1% of outstanding loans in a severe stress situation. In other words, it was only around a year ago, that RBI started to get it right on its estimation of NPAs.
While RBI’s latest FSR—where current NPAs of 10.8% will fall to 10.3% in March 2019—looks more credible since the latest Credit Suisse reports also talk of stress levels declining a bit, it is still early days and a lot can still go wrong. For one, NPA levels are still very high, and when borrowers have weak financials, NPAs can very well start rising again. Even now, Credit Suisse data shows that the share of total debt that is held by companies that have an interest cover of less than one is still a high 42%, though this is lower than 43% in the previous quarter. Within this, while real estate debt is as high as `5 lakh crore, this has been growing at over 20% per year over the last five years. Since most real estate developers have weak balance sheets—thanks to falling sales growth and high inventory levels—the ideal way to contain this is to be careful about real estate loans; yet, as was made clear in the government-RBI standoff, one of the government’s aim was to keep growing real estate debt.
In the case of telecom, where debt levels are around `3.25 lakh crore, Credit Suisse points out that all this debt is owed by companies with an interest cover of less than one. While firms like Bharti Airtel will try to ensure they don’t default, there is just that much these firms can do to avoid default if the government doesn’t move quickly on relieving their stress; though the stress has been building up for several years, there has been no concrete action so far on either lowering government levies or in lowering spectrum costs.
And while 35% of all steel debt is owed by firms that have an interest cover of less than one despite all the government protection, around 45% of power sector debt—of over `2 lakh crore—has an interest cover of less than one. In this case, too, while the government was leaning on RBI to change its norms to stop classification of these loans as NPAs, as this newspaper has pointed out earlier (goo.gl/MkpoYn), much of the problem lies in government policy; if this is not fixed at the earliest, and power loans are classified as NPAs, bank NPAs could soon start rising again.
Apart from the sectoral policies that the government needs to deliver on quickly if NPAs aren’t to start rising again—and fast—there is the issue of weak PSU banks that have, at present, lending restrictions under RBI’s Prompt and Corrective Action (PCA) policy. While the government wanted PCA norms to be relaxed so that these banks could start lending again—this was a big friction point with Urjit Patel before he resigned—as a recent
Kotak report points out, the gross NPAs of PCA banks were 21% of their total lending in the second quarter of FY19 (that for IDBI Bank was a whopping 31.8%) as compared to 11.5% for the non-PCA PSU banks. So if the government is able to get new RBI Governor Shaktikanta Das to relax the PCA norms—RBI’s latest Trends and Progress report suggests Das may not be in a hurry to relax these, but it is early days—that alone is likely to push NPA growth up.