Banks Sector rating: Cautious; PSBs: Underperform; Yes Bank, Axis Bank, ICICI Bank: Outperfrom We believe FY16 is going to be another choppy year where ROEs (returns on equity) are unlikely to improve materially. Falling NIMs (net interest margins), high credit costs, weak loan growth and high equity raising, particularly by the PSU banks, should result in earnings downgrades and poor ROE. We maintain our counter-consensus cautious stance on the banking sector and stick with private banks. Yes, Axis and ICICI are our top picks in the sector. We are 23% below consensus on average for our EPS (earnings per share) numbers on PSU banks for FY16e and FY17e. Capitalisation is the biggest challenge: Capital is now required not only for growth but also for solvency, in our view. PSUs remain heavily under-capitalised with CET1 (common equity tier 1) of about 8% and stressed assets at 13% of loans with 20% PCR (provision coverage ratio). The eventual CET1 under Basel-III is likely to be around 9.5-10%, in our view. Over the next four years, going by government estimates, PSU banks need to raise $35-40 bn. We believe the government will have no choice but to either recapitalise them or force large PSU banks to acquire the smaller banks. EPS downgrades to continue; bull case scenario unlikely: Since Sep \u201914, consensus has cut EPS by 14% and 11% for FY15e and FY16e respectively for PSU banks. This is despite bond yields falling by more than 125 basis points and PSU banks reporting large treasury gains. Private sector banks on the contrary have seen marginal EPS upgrades. We expect further earnings downgrades for PSU banks driven by lower loan growth, higher credit costs and equity dilution. For us to get constructive on PSU banks, we need credit costs to at least come down to 70 bps (vs. 130 bps assumed). Under this bull case assumption we see a 50% increase in EPS estimates and 25% upside to CMP (current market price). Credit costs\/NPL provisions will remain high: An under-provisioned banking system with just about 20% coverage on stressed assets for PSU banks is bound to see credit costs remaining high. The revised rules of RBI, which prohibits restructuring w.e.f. 1 April 2015, is also likely to exert pressure on credit costs. As a result even if stressed asset formation comes down, credit cost numbers won\u2019t and we expect credit costs for PSUs to be around 130 bps for FY16-17e, similar to the levels seen in FY15e. The private sector banks should see credit costs of 70 bps, again similar to FY15e levels. Rate cuts\u2013first-order impact is negative: The first-order impact of rate cuts is negative on margins as banks\u2019 liabilities are at fixed rate and assets are largely floating. During the previous falling rate cycle, PSU banks\u2019 NIMs declined by a whopping 50 bps in two years. Private sector banks on the contrary have maintained (or even improved) NIMs during rate cut cycles owing to their ability to manage ALM (asset liability management) better. We also don\u2019t expect significant benefits from treasury profits. RBI\u2019s Basel-III LCR (liquidity coverage ratio) rules will also keep a check on margins and ability to monetise SLR (statutory liquidity ratio) investments. Stick with high quality private banks despite valuation gap: We still prefer private sector banks over PSU banks. Even well-capitalised large PSU banks will likely be forced to acquire weaker PSU banks, which is very negative for shareholders of these large PSU banks. While PSU banks are trading at +1SD (standard deviation) below historic averages and private banks are at +1SD above historic averages, we expect that valuation gap to persist. Yes, Axis and ICICI are our top picks in the sector. We are Underperform rated on all PSUs.