The banking system is slowly recovering from the financial crisis. The number of banks failing each month has declined from high double digits in 2009 to only a few today. Many banks are returning to profitability and distributing their gains to shareholders: according to the FDIC, in Q4 2012 commercial banks and savings institutions distributed $35.5 billion in dividends to shareholders, up from $22.7 billion a year ago. As the economy slowly starts to pick up and credit improves, the growing consensus is that this trend will continue, especially among smaller, more conservative banks.
The market has taken notice, and Banking ETFs have outperformed the market so far this year:
SPDR S&P Regional Banking ETF (KRE)
Financial Select Sector SPDR (XLF)
SPDR S&P Bank ETF (KBE)
Vanguard Finacials (VFH)
While I agree with this perspective in the medium to long term for smaller and medium-sized banks, we are not totally out of the woods yet. There is at least one more major round of write-downs coming because banks have gotten a little ahead of themselves. In their eagerness to report net income growth and distribute dividends, banks have put off recognizing problems that still lurk on their balance sheets. As a result, reserves for loan losses will likely spike in the coming quarters, cutting into reported profits.
Reserving for Losses
Reserving for loan losses is always a mixture of art and science, because it combines current, objective balance sheet information with judgments about the future. Fair-minded people can disagree regarding the proper amount to set aside. The banking system as a whole can go through cycles of exuberance and caution in which rosier or more pessimistic views predominate. Much of the recent economic data has been good, including decent jobs numbers and rising housing prices, so it is not surprising that banks in general feel increasingly comfortable distributing gains to shareholders and are worrying less about delinquent or non-accrual loans. Reported profits for FDIC-insured banks in the aggregate are almost back to where they were in the bubble years:
This seems a little frothy to me, because there are still many balance sheet issues. In Q4 2014, 3.60% of total loans were noncurrent, compared to about 1.00% for most of 2004 and 2005 (Source: FDIC). Those bad loans are still clogging up balance sheets and many will eventually be written off. In the meantime, banks can "pretend and extend" by not realizing those losses quite yet, giving investors an inaccurate picture of their true health.
Modeling Future Reserves
To confirm my skepticism, I constructed a model that predicts provisions for loan losses in future quarters as a function of the following: last quarter's net interest income, non-interest revenue, non-interest expense, total net income, the ratio of current reserves to assets in non-accrual status, the ratio of current reserves to noncurrent assets, and the ratio of loans between 30 and 89 days late to loans 90 or more days late. These data are publicly available going all the way back to Q1 1984 and are available here.
With a sample size of 115 (data is quarterly and available beginning in Q1 1984), the model has an r-squared of 0.89. So here is what the model predicts:
Estimated loss provisions are more than $22 Billion next quarter, with a 95% confidence interval of $20.2 Billion and $25.6 Billion. The key driver here is that while reserves relative to non-accrual assets have risen to over 100%, loans between 30 and 80 days late remains stubbornly high and will weigh on balance sheets in the near term.
This matters because a low interest rate environment and slow recovery means that revenue has been pretty flat:
With minimal top line growth, a key driver for the bottom line in the last year was decreased provisioning for losses. When banks are finally forced to recognize their recent aggressive accounting, many of these paper gains will be reversed.
How to Play This
While I like some financial stocks, these findings indicate that caution is warranted. Many banks still trade at attractive valuations, although investors should examine reserves and recent net income numbers with a skeptical eye. Investors wanting exposure to banks may be better off selecting individual names rather than sector ETFs because the wider diversification of ETFs means that they will likely be caught in the coming storm. I personally like BofI Holdings (BOFI) because it is fast-growing and has very conservative underwriting. I make a more detailed case for this stock here.
For those wishing to maintain long positions in bank ETFs, the recent low volatility environment means that options are selling at attractive prices. Because balance sheet problems will likely become apparent in the spring or summer earning seasons, low cost portfolio insurance can lock in recent gains and protect against market declines. For protection through the spring earnings season, the Jun 13 31 puts on KRE cost $0.81. At 2.5% of the current cost of a KRE position, this hedge maximizes potential losses to 5.2% including hedging costs. For protection across both the spring and summer earnings seasons, the SEP 13 31 puts on KRE cost $1.42. At 4.4% of the current position, losses through September are maximized at 7.1% including hedging costs.