The Third First Savings and Loan corporation is trading at a significant discount to its intrinsic value exists because generally accepted accounting practices overstate the number of economically relevant shares to be used in determining common financial ratios. As a result, the thrift appears more expensive at first glance than it does upon closer inspection. Management recognizes the opportunity this is creating and is aggressively buying back stock at highly accretive prices. These purchases along with the upcoming dividend reinstatement should help TFSL trade closer to fair value.

If you look up Third Federal Savings and Loan Corporation (TFSL) on any online data service, the market cap will be reported as $4 billion, price to book at 2.2, and a share count of 304 million.   Yet, of the 304 million TFSL shares on file with the SEC, only 73 million shares at the corporation are economically relevant at this time.  The remaining 227 million are in reserve until the demutualization process completes.  Since these share have yet to be sold and do not receive dividends, I believe they should be treated as dormant treasury shares until their full value is realized and the capital received in their exchange is placed on the books. Adjusting the share count drops the market cap of TFSL to 1 billion. It also places its price to book ratio at a tantalizing 57%. To explain the justification for this accounting change, a little background on Third Federal and the two-step demutualization process is necessary.

The Third Federal Savings and Loan Corporation (TFSL) is a mid-tier mutual holding company which owns the Third Federal Savings and Loan Association. The association operates branches in Ohio and Florida with over 11 billion in assets. It was founded in 1938 by Ben and Gerome Stefanski. Until its 2007 IPO, the ownership of the mutual was 100% owned by the members, that is the depositors, of the thrift.

In the early 20th century, mutual companies were common form of organization for businesses, especially savings thrifts and insurance companies.  Investors who are fans of the movie, “It’s a Wonderful Life” will recognize the Bedford Falls Building and Loan as a mutual association. In the film, the townsfolk’s deposits weren’t just their savings, they were also shares in their mutually owned Building and Loan. Of course, that’s why old man Potter wanted to buy their deposits for fifty cents on the dollar, he wanted to gain control of that miserly little old institution.

Because the depositors of a mutual thrift are also the owners, tracking who owns what can be a little confusing when holding companies are involved. Following is a list of the three different organizations named Third Federal Savings and Loan and their distinguishing features:

Third Federal Savings and Loan Mutual Holding Company (MHC)
(74% owners of the Third Federal Savings and Loan Corporation)
(100% owned by the members of the Third Federal Savings and Loan Association) 

Third Federal Savings and Loan Corporation (NASD: TFSL), the mid-tier holding company which owns 100% of the TFSL Association.
(26% owned by public shareholders, 74% owned by the MHC)
(100% owner of the Third Federal Savings and Loan Association)

Third Federal Savings and Loan Association, a regulated thrift institution
The members (depositors) elect the board of directors to the MHC which maintain majority controls of the corporation for the benefit of the association.

Mutual thrifts depend on their retained earnings and new deposits for growth capital. This leads to steady but slow growth which probably suites most members of the association just fine, but not necessarily their more ambitious managers who usually want to grow as fast as their peers at stock owned corporations.

To accommodate a mutual’ s need or desire for more capital, the Federal Reserve permits them to convert to stock holding corporations following well established procedures and regulations.  Depending on how much capital they seek, they can chose to either partially or fully convert from mutual ownership to a stock corporation. In a partial sale, up to 49%, the mutual association gets to raise capital while retaining control of the thrift, while in a full conversion, more capital is raised but the resulting corporation is now solely in the hands of the stock holders.

In a one-step conversion, shares in the new corporation are offered first to the current members of the association. The offering is usually at an attractive price recommended by an “independent appraiser”.  Frequently, the appraised price doesn’t fully reflect the intangible factors of value such as the mutual’s branch network, relationships, depositor base, and especially its increased value as an acquisition target. Not surprisingly, the share subscriptions are almost always oversubscribed.  Since the amount of capital raised in a one-step conversion could be more than its existing retained equity, a single step conversion can easily raise more capital than the thrift can effectively manage.

As a result, most mutuals chose to convert in a two-step demutualization process. In step one, the thrift is transferred to a mid-tier holding company that issues the shares on behalf of the association.  Up to 49% of the shares are sold to the members and sometimes the general public to raise capital, while the remaining, majority of the shares are assigned to a new mutual holding company (MHC).  Because the public shares are a minority position, they are often referred in financial analysis as the “minority shares” of the new corporation.

The board of the MHC is elected by the members of the association from a slate of nominees provided by the management. The MHC board then controls the mid-tier holding company which owns the thrift. The result is that capital can be raised by the mutual without its members or managers losing control.   Should later, the thrift want to raise more capital, or go fully public, it can then sell the rest of the MHC shares to its members and become 100% shareholder owned.  Most thrifts chose to convert in two steps rather than one to better manage the amount of capital they raise at any one time.

Some thrifts, including TFSL chose to only complete the first step of a two-step conversion and remain indefinitely at step one.  This allows them to obtain more capital, but remain with the mutual holding company structure preferred by the members and management.  This is in contrast to most thrifts where the management seeks the glory and gain of selling to a larger public institution at a premium and encourages the mutual membership to complete a second-step a few years after the first step.

In the case of TFSL’s step one IPO, 30% of the corporation was sold to the public raising $886 million including a $106 million dollar loan to the ESOP which bought 5 million in public shares with the proceeds, 1.5% was granted gratis to the Third Federal Foundation, and the remainder was issued to a newly created Third Federal Savings and Loan mutual holding company (MHC).

The board of TFSL’s MHC is elected by the members (depositors) of the Third Federal Savings and Loan association by voting to approve a slate of nominees provided by the bank’s management.  The MHC structure allows the association to maintain majority control of the corporation while raising capital from the benefit of the savings association.  While independent nominees can be selected, this never happens, as the depositors are far less interested in these details of the mutual than is the management.

Should TFSL ever fully demutualize, the MHC would dissolve by selling its shares of the corporation to the members of the association in a second step conversion.  The capital raised by selling the MHC shares would then drop down to the association creating an extremely, if not excessively, well capitalized thrift.  Since the capital raised by the share subscription stays within the corporation, the members who take advantage of a second step conversion are effectively paying for their shares with their own capital and getting a portion of the retained earnings for free.  Hence, most subscriptions are oversubscribed. Upon completion of the second step, the new corporation would be completely shareholder owned, over-capitalized, and led by a management that would quickly become more profit focused than before. The result is a highly attractive takeover candidate.

In our case, TFSL is showing no interest in completing a second step conversion. Mark Stefanski, 59, CEO and Chairman of the board, started at the thrift in 1981 and took on his current roles in 1988 at age 34 when his dad, the original CEO retired after 50 years of service.  While he grown the bank five-fold in his 25 years, his current returns on equity and assets are unimpressive given the thrifts overcapitalization, though the numbers are trending favorably out of the recession. Besides being very well compensated, Mark Stefanski is leading a bank with a unique culture and it is hard to envision he would want to leave or change it.

While Mr. Stefanski is unlikely to be confused for John Stumpf at Wells Fargo for profitable growth, shareholders can be thankful that he does solidly understand the value of properly priced buybacks. Since 2007, TFSL has bought back 29 million shares of the publically traded shares. The public float is now 24%, or 70 million shares, down from 105 at the time of the step one IPO.

There is a caveat to ignoring the MHC shares in financial calculations.  When (if) a mutual fully converts, and the shares at the MHC are sold, it is almost a certainty that they will be sold to the members at an attractive discount to their true value, as evidenced by the 10-20% jump on the IPO most conversions experience. This discount is dilutive to the public shareholders, but they too gain from the IPO jump, so some of the dilution is gained bank.  However, a few percent of the shares are often given gratis to a local community foundation or as well as a usually larger stake that is given to establish a management incentive pool. Those shares will be 100% dilutive to the public, but the second step conversion also makes the new corporation an extremely well capitalized and desirable takeover target which is accretive to the public shareholders.  The net effect of these actions will likely be a small, but manageable dilution for the public holders when the second step occurs, but it does not break the investment thesis.  In any case, at TFSL, a second step is probably years away.

Financial Ratios based on adjusted share counts

TFSL Corporation Shareholder equity 1,864,864,000
August 4, 2014 SEC 10-Q Share count 304,096,983
Shares held by MHC 227,119,132
Public (Minority) shares 76,254,395
Current Share Price 13.60
Share Count Market Capitalization Book Value P/B ratio
MHC + Public Shares 4.1 Billion 6.147089 2.21
Public Shares only 1.0 Billion 24.45582 0.56
Dates Buyback shares purchased Average cost
2008 – 6/30/2014 29,444,050 est. 11-12
4/19/2014 – 6/30/2014 3,144,050 13.43

Third Federal’s growth strategy

Third Federal Strategy for growing value is based on three elements: growing the mortgage portfolio with adjustable rate and 10 year fixed loans to protect exposure to interest rate risk, dividends, and buying back the minority shares.   Of these, the buy-back of the public shares is the surest path to success given the accretive value of buying shares in a financial at .56 book.  The company is currently purchasing around 60,000 shares/day or close to the legal maximum shares allowable of 25% of the daily volume on a 5 day moving average.   On the August conference call, management said there is nothing “standing in the way” to prevent them from authorizing a new buyback program as soon as this one completes.  Given, how clearly management has reiterated its commitments to buybacks, this is almost a certainty.  However, there may be a 30-45 day delay between filings while awaiting approvals or “non-objection” from the Fed.  Interestingly, this delay might be an optimum time for individuals to buy more shares as the company’s buying pressure will be missing.

All this buyback talk begs the question.  Where did TFSL get all this capital to fund its buybacks?  The answer is that TFSL is still grossly overcapitalized from its IPO in 2007.

Third Federal Savings and Loan Association Actual Ratio   Required Ratio
Total Capital to Risk-Weighted Assets $1,648,098,000 23%   $710,810,000 10%
Core Capital to Adjusted Tangible Assets $1,565,596,000 13%   $583,896,000 5%
Tier 1 Capital to Risk-Weighted Assets $1,565,596,000 22%   $426,486,000 6%

June 30, 2014

TFSL Corporation   Actual Ratio
Total Capital to Risk-Weighted Assets   $1,942,581,000 27%
Core Capital to Adjusted Tangible Assets   $1,860,079,000 16%
Tier 1 Capital to Risk-Weighted Assets   $1,860,079,000 26%

From 10-Q, June 30, 2014

TFSL successfully preserved its capital during the financial crisis of 2008-9 by staying with mostly vanilla loan products, it did not seek liar loans, employ commission driven mortgage brokers, or act with reckless abandon.  However, it did have a small community reinvestment program (about 1.2% of all loans) for residential mortgages called “Home Today” which applied “less stringent underwriting and credit risk standards”. Approximately half of these loans used private mortgage insurance, but the majority of those policies were underwritten by PMIC, which is now in receivership and expecting to pay out only 67% of its claims.

These loans are stinkers. Though small in number, 5 years later this tiny pool of loans is still responsible for nearly 25% of all 90+ days’ delinquent receivables. The good news is the bank fundamentally changed its community reinvestment loan program in March 2009. Borrowers without sufficiently large down payments who qualify for the “Home Today” program must now meet the same minimum credit scores and its new private mortgage insurer requirements as standard borrowers.

The other trouble spot during the financial crisis was poorly underwritten home equity loans or second mortgages.  Currently, 16% of the thrifts loans are home equity loans and 18% of the company’s 90+ days delinquent receivables are attributable to these loans.  In June 2010, the thrift made substantial changes to its home equity loan program and currently writes only a fraction (5-10%) of the volume in 2013 than it did in 2007-8.   As the 10-K states, “When the Association began to offer new home equity lines of credit again, the product was designed with prudent property and credit performance conditions to reduce future risk.” Translated, this means the thrift writes only a small fraction or 5-10% of the home equity loan volume today than it did in 2007.

The quick rise in second mortgage delinquencies, not only generated quick losses, it also raised the ire of the Office of Thrift Supervision (OTS).  Upon closer examination, OTS issued a memorandum of understanding (MOU) to the MHC in February 2011 suspending any the issuance of new debt, dividends, or share buybacks until enterprise risk controls were improved. The risk control MOU was not lifted until April 2014.

The Importance of Book Value

Investors will argue what P/B should be paid for a thrift until the end of time, but just about everyone agrees that for a profitable thrift with nearly 2 billion in equity and a  good chance at earning average ROE’s,  56% of book is cheap. Certainty, if TFSL were to complete its second stage conversion, it would sell for more than book value in its takeover.

In particular for mutuals completing a second step, the book value and the price of the public shares ares often the key metrics used to compare the firm with its competitors when determining at what price the MHC should sell its shares to the association’s members.  Since this metric will determines how much capital is raised in the second step conversion, it is clearly in the management’s best interest to increase it.

So how much stock can TFSL buy back and what will that do for the remaining shareholders?

Currently, Total Capital to Risk Weighted Assets is 1.94 billion or 27%, if it was dropped to 1.4 billion or 20%, TFSL would still have twice the capital required and free up $500 million for share repurchases.  The only real limitation is how many shares a day it can buy in a day.  With 250 trading days in a year, fifty thousand shares a day would allow for 12.5 million shares to be purchased.

Public Share Count            76,254,395
Current Equity  $  1,864,864,000
Capital available to spend on buy-backs  $      500,000,000
Equity after buy-back  $   1,364,864,000
  Scenario A Scenario B Scenario C
Average buy back price $                     15 $            16 $            17
# of shares purchased 33,333,333 31,250,000 29,411,765
Buyback Timing @ 1Mil shares/month 33 31 29
Post Buy-Back Public Share Count 42,921,062 45,004,395 46,842,630
Reduction in public shares 44 % 41 % 39 %
Book Value $               31.80 $      30.33 $      29.14
P/B Valuations Post buyback price at different valuations
  50%  $               15.90  $      15.16  $      14.57
Current 56%  $               17.81  $      16.98  $      16.32
  66%  $               20.99  $      20.02  $      19.23
Reasonable 75%  $               23.85  $      22.75  $      21.85
  85%  $               27.03  $      25.78  $      24.77
Takeover Price 125%  $               39.75  $      37.91  $      36.42

So what kind of return can TFSL common stock generate?

Using scenario B, book value would rise to $30.33 from the reduction of shares below book value as 31 million shares would be purchased over 2.5 years. The thrift is likely to retain approximately 3 million per month after dividends or 90 million over these 30 months which would add $2 to book value (not included in above chart). Additionally, 70 cents in dividends would be received. The return on a $13.60 investment would be 14.1% per annum if the stock continues to trade at 56% of book, but at a more reasonable 75% of book, the return would be 28% per annum.

I don’t consider an increase in the price to book value from 56% to 75% to be unreasonable for a profitable mutual demonstrating a commitment to returning capital to its public shareholders.


A note about mutual holding company dividends.  MHC don’t want the dividends.  Accepting dividends generates a taxable event.  The most cost effective to get more money into the hands of the association’s members isn’t through dividends, but by paying higher interest rates to the depositors who are the members. Of course, the public shareholders do want the dividends, so the traditional MHC answer is to waive their right to the dividends.

Prior to the Dodd-Frank act, the board of the MHC could waive the dividends issued by a mutual thrift by a board decision. Post Dodd-Frank however, MHC’s are required to have its members vote on the waiver decision.   On July 31, the members of the association, voted to approve the board’s recommendation to waive up to 28 cents of dividends for the next year by a 97% approval of those who voted. The request to begin the dividends at seven cents per quarter is now at the Fed awaiting a non-objection notice for up to 45 days. Given the strong capitalization of the thrift, the lifting of the MOU earlier in the year, the previous Fed approval of the earlier buyback, and the dividend coverage from earnings, its approval seems a formality.

A 28 cent dividend equals a 2% yield at the current $13.60 share price. Given the aggressive decrease in share count, the dividend payout per share is also likely to rise quickly as the dividends from the retired shares are directed to the remaining shares.  If half the stock is retired in 5 years (not an unlikely scenario given the company’s history), the dividend could double without a change in the aggregate dividend amount paid to shareholders from the company.  Dividend coverage is about 30% of earnings.


The biggest risk to TFSL, is a sudden increase in interest rates.  The thrift has moved over the past 4 years to lower its percentage of long-term fixed mortgages that it retains from 59% of the outstanding loans to 39% by focusing on ARMS and 10 year fixed mortgages.  Still, an increase in interest rates would definitely hurt.  As of June 30, 2014, in the event of an increase of 200 basis points in all interest rates, the Association would experience a 16.05% decrease in equity.   Your guess about interest rates is as good as mine, but I am pleased that the management is continuing to reduce long-term exposure.

The next unknown is the final implementation of the Dodd-Frank regulations concerning mutual holding companies.  For example, MHC’s are now required to obtain member approval to before waiving dividends. Will other restrictions be placed on capital transactions?  Compliance costs are already up, what’s next?

The company froze its defined pension plan in 2011.


Current bank performance

I haven’t emphasized the thrift’s performance, because all the thrift needs to do to make my investment thesis work is avoid big mistakes, but a review of some key numbers are in order.  First, the thrift did maintain profitability through the recession and all ratios except expense are trending favorably.  While the return on assets at .61% and a return on equity of 3.75% are completely uninspiring, deleveraging the balance sheet through buybacks will increase both materially as will the continuing decline in delinquent loans.   Lastly, expenses are higher and with increased regulation, and  I expect them to stay elevated for the foreseeable future.

  30-Jun 30-Sep 30-Sep 30-Sep 30-Sep 30-Sep
  2014 2013 2012 2011 2010 2009
 Selected Financial Ratios and Other Data:            
 Performance Ratios:            
 Return on average assets 0.61% 0.50% 0.10% 0.09% 0.10% 0.13%
 Return on average equity 3.75% 3.05% 0.64% 0.53% 0.65% 0.80%
 Interest rate spread(1) 2.24% 2.25% 2.11% 1.97% 1.77% 1.70%
 Net interest margin(2) 2.41% 2.46% 2.39% 2.32% 2.16% 2.20%
 Efficiency ratio(3)   59.81% 59.67% 60.31% 56.59% 54.59%
 Noninterest expense to average total assets   1.58% 1.52% 1.54% 1.50% 1.51%
 Average interest-earning assets to average interest-bearing liabilities 118.49% 119.58% 119.60% 120.39% 119.70% 120.57%
 Dividend payout ratio(4)   525 % 520%
 Asset Quality Ratios:            
 Non-performing assets as a percent of total assets(5)  1.19% 1.58% 1.76% 2.34% 2.73% 2.57%
 Non-accruing loans as a percent of total loans(5) 1.32% 1.53% 1.77% 2.37% 3.08% 2.74%
 Allowance for loan losses as a percent of non-accruing loans(5)  59.08% 59.38% 55.03% 66.73% 46.49% 37.33%
 Allowance for loan losses as a percent of total loans(5)  .78% 0.91% 0.97% 1.58% 1.43% 1.02%
 Capital Ratios:            
 Total risk-based capital (to risk weighted assets) 23.19% 24.10% 22.19% 22.29% 19.17% 18.19%
 Tier 1 core capital (to adjusted tangible assets) 13.41% 14.18% 13.31% 13.90% 12.14% 12.48%
 Tier 1 risk-based capital (to risk weighted assets) 22.03% 22.85% 20.94% 21.04% 18.00% 17.30%
 TFS Financial Corporation(6)            
 Total risk-based capital (to risk weighted assets) 27.24% 27.94% 25.03%  NA  NA  NA
 Tier 1 core capital (to adjusted tangible assets) 15.89% 16.59% 15.33%  NA  NA  NA
 Tangible capital (to tangible assets)   16.59% 15.33%  NA  NA  NA
 Tier 1 risk-based capital (to risk weighted assets) 26.08% 26.69% 23.78%  NA  NA  NA
 Average equity to average total assets 16.21% 16.38% 16.00% 16.07% 16.19% 16.69%
 Number of full service offices   38 39 39 39 39
 Loan production offices   8 8 8 8 8


Disclaimer: The author owns TFSL common.