Description

Summary:

TFS Financial, TFSL, is a Cleveland OH based bank that is the largest mutual holding company in the market.  Its public stock trades dirt cheap, at 8.8x runrate earnings and at 45% of book. Over the past 5 years, TFSL has typically traded in the 12-15x earnings range, pretty typical for a conservatively run regional bank.

But since late 2016, TFSL has suffered despite decent earnings.  Different from many banks, TFSL’s NIM compresses slightly as the Fed raises rates, and slower earnings growth has pushed the stock down to its lowest valuation multiple in years.  It has been written up a couple of times on VIC before, as it has had some regulatory bumps along the way, which all have been resolved awhile ago.

Today however, it seems a name that has been left for dead as it lacks any catalyst (not likely to do a second step mutual conversion, nor can it be acquired).  That said, the stock trades at a ridiculous 6.75% dividend yield, an 11% FCF yield, and likely will buy back well over 10% of the public float in the next couple years.  Buybacks, loan growth and a huge dividend should enable the stock to break out of its doldrums. Over the next 2-3 years, this stock could easily trade back to a 10-11x multiple, or around 21-22 per share.  With dividends, investors can easily earn 15-20% annual returns on their capital over the next 3-5 years.

Financials

Mutual Holdco (MHC)

I hate to get too much into the weeds here, as other write ups on VIC have discussed the structure here, but basically TFSL is a second step thrift conversion bank.  Basically, there is a dual ownership structure, with 227mm shares held by Third Federal Savings who have voting rights, but no economic interest in the bank, plus the 53mm publicly traded shares out today.

TFSL originally sold 100mm shares to the public in 2007 at $10 per share in the so-called first step.  (Highly accretive buybacks have reduced the share count by almost half in a decade).

The second step occurs when the MHC shares are issued to the public.  That obviously hasn’t happened yet, and based on CEO comments, seems quite unlikely for some time.  I deem the MHC shares really as authorized but unissued shares. Depositors, who would have the rights to buy these shares in a second step conversion, would have to pay for these shares, and the bank would obviously get cash.  Accounting rules however requires financial statements to report EPS and share data using the full 280mm share count. Obviously this is quite misleading.

Second step conversions are usually done I’m told between 80-85% of fully diluted book value.  A second step here would mean the company issuing around $3.5BB of equity, a huge amount for a bank with a float of $800mm today.  Very acquisitive mutual banks, or ones that need capital for regulatory or other purposes, are ones much more likely to take this second step.  The bad news is, this can keep a lid on the equity value for the public shares. The prospect of dilution, as well as a dividend that would in all likelihood be turned off for awhile clearly would be an overhang.

That said, even if TFSL did a second step at 82.5% of fully diluted book, that would imply selling the MHC shares at $15.50, well below where TFSL is trading now.  With the company then demutualized, it would likely then trade to 1.1 – 1.2x book or better depending on the markets view of the underlying business and the comps of course.  (Kearny Bank traded quickly to 1.2x book post its second step). Just 1.1x book value for TFSL proforma for a second step would take the stock to $17.05 per share, 15% higher than its current price.  This is a downside case however in my view.

More likely, TFSL, at 20% Tier 1 capital and not terribly interested in making any acquisitions, looks well overcapitalized.  Marc Stefanski, the son of the co-founders (his parents), has been running the bank for over 30 years, and seems to more intent on growing organically, and leaving the decision for a second step conversion “to the next generation.”  He also, at 63 years old, recently said in an interview that he doesn’t know what the word retirement is. He could be running the bank for 10-20 years more.

Business

TFS Financial is a straightforward lender to consumers for mortgages and home equity loans.  They are based in Cleveland and have 29 branches in Ohio, and 17 in Florida.

Their loan book today is $12.6BB:

-          47% First lien fixed mortgages, $5.9BB

-          40% First lien Adjustable Rate Mortgages (ARMs) or roughly$5.1BB, and

-          13% in home equity lines of credit, $1.6BB

For the first lien mortgages they own, the average FICO score is a healthy 772; LTV’s are at 69%, and delinquencies are extremely low at $12mm out of $10BB in loans.  There is some legacy subprime loans they did before the Great Recession (via their Home Today program), but these are mostly worked out and loan losses were sub $1mm in FY 2017.

This below is from the last 10Q, and really tells the story.

Loan yields are low as they underwrite conservative mortgages.  Many of its HELOC’s are underwritten to its existing mortgage customers.  They generally lend long, and borrow short. Demand deposits are obviously a great means of funding their loan book – CD’s are only ok but can help manage rate risk as they are laddered and roll off gradually.

The problem I see with the balance sheet is that they fund too much of their longer dated mortgages with CD’s (many brokered), and FHLB borrowings (the “Borrowed funds” above).  The FHLB loans adjust quickly to higher rates, and while ARMs also adjust relatively quickly, fixed mortgages of course do not. Their longer dated mortgages of course are negatively convex too, and so duration there increases as rates rise.  That means Fed raises, especially rapid rate increases not offset with natural loan attrition (via sales, refi’s or moves), can be a negative to TFSL’s net interest margins (NIMs).

Indeed, NIMs have fallen from 2.2% in FY 2015, down to 1.93% in the last quarter.  Not a disaster, but compression nonetheless. Further Fed increases, unlike many depository franchises heavier in checking and savings accounts, will probably push NIM’s down a bit further.

Recent Results

I typically try to look past quarterly numbers, especially when one-time items create selling pressure.  In Q2 for example, TFSL celebrated its 80 year anniversary with company-wide events in both Florida and Ohio.  According to management, the celebrations cost $2mm (what?!), and another $2mm was spent this year on bonuses (a $1000 per employee bonus to all employees this Q2 as well).  Their non-interest costs have jumped, and earnings are comping flat to down slightly.

Still, over the past few years the company has kept a lid on non-interest expenses, and 80 year anniversaries don’t happen every day or year.

Below is the income statement from the past 5 years.  

The normalized runrate column on the far right uses a 20% tax rate, and excludes $2mm of costs from the June quarter.  (Because their FY end is September, the tax rate is blended over every quarter in the current fiscal year at 25.5%, but the go forward number will be 20%).  I also excluded the credit for loan losses. That gets EPS to $1.68.

After the Great Recession, losses on their home equity line of credit (HELOC) portfolio appeared severe.  Generally, these lines allow homeowners 10 years from which to draw on the LOCs, at which point they converted into 10 year amortizing loans.  There was no amortizing portion of any draws in that 10 year period. TFSL well over-reserved for these in 2011-2012, and now is reversing those allowances.

So, too high provisions became credits, and have actually boosted earnings since 2015.  Recoveries on delayed draw HELOCs ended up being far better, and delinquencies far lower than management anticipated.  While my history is short with this name, probably the MOU fight with the Fed back in 2012-2013 caused the company to be extremely conservative in their reserving for future losses.

In any case, the 10 year HELOCs issued back in 2008-2009 (some extended a bit), still will roll into amortizing loans in 2018 through 2020.  As of June, HELOC loan loss allowances totaled $21mm, but there were only $5.9mm of loans over 90 days delinquent. Likely recoveries will outpace charge-offs given home price appreciation, so I expect another $15mm in reserve releases just related to this book in the next 2-3 years.  (There is a schedule of when the 10 year draw periods end in the 10K, and this is when the company re-evaluates reserves). That should provide a nice cushion to provision for loan losses in the next couple of years. Conservative underwriting and low unemployment should also help loan loss figures in the near to intermediate term.

Projections

It’s pretty easy to come up with scenarios for earnings, but clearly impossible to get right.  I generally took a very conservative perspective over the next few years, to help underwrite the potential downside.  My net income numbers in fact are 10% below street figures. Partially I do think the street is too high, but also I normalized loan losses to capture a more accurate view of earnings (despite my view that loan reserve releases probably continue).

The key inputs here are loan growth and projected NIM’s.  Chargeoffs over the past couple of years have averaged around 13mm per year, and is a decent base for what provisions for loan losses will be.  The share count will eventually fall by the 6.6mm shares left under their buyback plan, probably at a slower pace than in the past, but likely to be completed over the next 2-3 years.

Management has been frustrated that buybacks have not done much for the stock, and instead of huge buybacks, are now focused on loan growth and the dividend.  Which is a bit unfortunate given the depressed valuation today. But no matter what, a lot of capital is being returned to shareholders.

This July, MHC shareholders approved a $1.00 per share dividend through July 2019 (an increase from $0.68 in the prior 12 months).  Seemingly, the market has completely ignored the 47% dividend increase. MHC agreed to waive any dividend for the fourth year in a row.

Valuation

I didn’t put together a list of any specific comps.  But I think it’s pretty straightforward to judge TFSL by its ROE’s, loan book and history of underwriting.  The litmus test in my mind is always, did the company lose money in 2008-2010, and how much? TFLS certainly had a big drag on earnings with their Home Today program (discontinued, permanently), but stayed above water and today the company maintains solid underwriting standards.  Loan losses and delinquencies remain very low, and I expect TFSL to continue making money in the next recession.

As for upside, I modeled out a bit over $2 in EPS in 2021, even with 20 basis points of additional NIM compression.  The upside case is one that seems not even a big stretch – that NIM’s return to 2%, their loan book grows at 4% per year, buybacks take out the remaining shares under their authorization (6.6mm shares remain authorized).  That would take EPS to $2.65 in 2021.

Some scenarios:

Sometimes I think investors forget the power of compounding dividends.  As an example, I own a closed end fund (PCI) and have for about 5 years.  The price is only up 7% cumulatively in that time, but with an 8% yield plus some special div’s, it has compounded at 13% per year.  Not bad.

The bank index today trades at 12x, and given the more expensive liability funding compared to other banks, I think 10x is fair for TFSL.  ROE’s are in the 5-6% range. Over the past 5 years, TFSL has grown BV/share from $25 to $33 roughly. It has traded in a range of 45-65% of this book too (excluding the MHC shares of course).  At the current discount, at 45% of book, it is at a valuation low when most banks’ valuations have improved dramatically. Said differently, book value has grown $8 per share in the past 5 years, but the stock has moved less than $3 per share.

Risks

NIM’s, delinquencies/defaults, economic conditions particularly in Florida and Ohio.

The CEO Stefanski has said in the past that making a profit isn’t the primary reason he runs the bank.  There clearly is a well-treated, happy workforce here (turnover is 3%). That has its own benefits, but clearly has had a recent impact on earnings.

Long duration portfolio.  Rate increases by the Fed will increase interest costs faster than their mortgages will refinance at higher rates.

Second Step Conversion.  A second step could mean the dividend goes away as well as earnings dilution.  A well-capitalized bank with a fresh $3BB+ in cash would probably struggle to deploy that quickly.

Catalyst

Earnings, dividend increases next year, improved financial performance