TFS Financial ($TFSL) — Initiation (7/26/20)

Price: $14 Market cap: $740m Valuation: 0.45x price to book Category: GENERAL

TFS Financial is the bank holding company for Third Federal Savings & Loan. This is a small-cap plain-vanilla bank.


The stock has been unimpressive over the years but had some good runs during various periods…

This will be an incredibly unexciting pitch but here it goes — it is a cheap stock with a solid dividend. That’s it!

Well there’s obviously a bit more to it than that but that’s the basic gist of the bet… In fact, I think the dividend alone could outperform the overall market for years from here.

Let’s start with some basics…


This is a savings & loan focused on residential mortgages in primarily Ohio and Florida. The loan book is as plain as it gets — residential mortgages and home equity lines of credit. (They’ve stepped back somewhat from HELOCs since the financial crisis.)


Mortgages are split about evenly between fixed rates and adjustable rates.


Third Federal is a former “mutual conversion” where depositors owned the bank and then took the company public — but not entirely. It is what’s known as a two step conversion with the second step not yet having been completed.


The best way to describe how mutual conversions (or thrift conversions as they are known) is to quote Seth Klarman from Margin of Safety:

In a typical initial public offering (IPO) all pre-offering shares are owned by insiders, who typically bought in earlier at a small fraction of the offering price. To illustrate the diluting effect of a typical underwriting, if insiders have bought one million shares of XYZ at $1 per share and the public is subsequently offered one million newly issued shares at $11 each, there will then be two million shares outstanding with total proceeds to the company of $12 million (ignoring underwriting costs). The pro forma book value of the company's stock is $6 per share. The public's investment has been diluted by $5 per mare (45 percent of the purchase price), while the insiders have gained a windfall of the same $5 per share. Thrift conversions work differently. A thrift institution with a net worth of $10 million might issue one million shares of stock at $10 per share. Again ignoring costs of the offering, the proceeds of $10 million are added to the institution's preexisting net worth, resulting in pro forma shareholders' equity of $20 million. Since the one million shares sold on the IPO are the only shares outstanding, pro forma net worth is $20 per share. The preexisting net worth of the institution joins the investors' own funds, resulting immediately in a net worth per share greater than the investors' own contribution.

With that background, it’s important to note that 227m of the 280m shares outstanding are owned by the Mutual Holding Company (MHC) — this means there are truly only 53m shares outstanding today.

The financial crisis did some damage to this business —

  • Roughly 1/3 of the loan book was in home equity lines of credit ($2.9bn on $9.3bn in total loans)
  • Today, HELOCs are about 17% of total loans — $2.3bn on $13.6bn in total loans
  • Loan losses went from $6m in 2005-2006 to $115m in 2009
  • They remained profitable in 2008 and 2009 and did not need to raise capital

Financial Review

(TFSL has a September yearend)

  • Loans have grown at a 4% annual rate from 2015-2019
  • Deposits grew ~1.4% per year
  • Equity has mostly been flat — thanks to dividends and buybacks
  • Earnings haven’t really grown but they’ve held steady
  • Dividends ratcheted up considerably — 14.5x from 2015 to 2019
  • One potential concern — they’ve been releasing reserves for years now… perhaps they legitimately over-provisioned during the credit crisis but it’s worth noting — remember in 2008-2009 they reserved $35m and $115m respectively, and now have been releasing about $10m per year
  • ROEs hanging between 4-5% annually — more so a sign of the overcapitalization and low(er) returns on residential mortgages
  • Net interest margins are LOW — TFSL has been at or below the 2% threshold for a while now… again, residential mortgages as your core focus will do that…
  • Plenty of equity — capital levels are excellent


TFSL has been busy with buybacks over the past 7 years — shares started at 309m in August 2013 and as of March 2020 they were 280m — this is deceiving due to the MHC status…

227m shares belong to the MHC and are technically not outstanding — this means the true share count went from 82m to 53m!

So we have 53m shares x $14 stock price = $742m market cap

Net income over the last 12 months was $82m or $1.55 per share = 9x earnings multiple

Book value is $1.65bn at 1Q20 or $31 per share = 0.45x price to book

We also have an extremely overcapitalized bank with $1.65bn in equity supporting $15bn in assets or 11% equity-to-asset ratio — compare this to the top banks that are well under the 10% mark.


Earnings per share have more than doubled from 2013 to 2019 and book value per share grew from ~$25 to $32 in that timeframe.

A lot of this growth came from the 50m+ shares repurchased since the initial IPO in 2007.

This is a great setup — fundamentals improving but stock price has gone nowhere.

Banks are complicated from a valuation standpoint… It’s easy enough to say they’re cheap on an earnings or book value basis. Cash flows from banks are difficult to estimate for starters.

Dividends are typically less difficult to estimate and (generally) less volatile — making the dividend discount model pretty useful here. A good rule of thumb equates the ROE of a bank to the appropriate book value multiple… For example, if a bank is earning 10% ROEs, then generally it should trade at 1x book value which would also equal 10x earnings.

Here’s my attempt at a basic DDM for Third Federal —


The way it works — assumes you get your dividend each year (discounted) and then in the final year (2031) all of book value is paid out… So book value is the final discounted cash flow.

My estimate assumes a 4.5% ROE for the entire period and a 10% discount rate. You can see most of the value comes in the large dividend payout ($7 in discounted value over 10 years). After which you’re still left with a bank having $35+ in book value.

The end result was $19 per share gets you an implied 10% return and would equate to ~0.6x book value today. This price target would imply a 6% dividend yield.

This seems about right… You have a large payout ratio which gets you most of your annual return but the company is still retaining some amount of capital to buyback stock or grow the loan book.

Final Thoughts

This isn’t a sexy growth bet… It’s more of a “punt on third down” bet… But it’s a pretty safe setup (even with the banking sector risk on interest rates). Getting the large dividend (8% yield) with a book equity cushion is nice. I’d be happy with the 8% return alone over the next 10 years.

One other facet I neglected to mention — the MHC structure is still there for use. Third Federal could decide to raise additional equity in the second step mutual conversion (not dilutive to current shareholders) and have even more capital. This adds a little bit of a special situation element to the stock and gives shareholders another arrow in the quiver to protect downside. I find it unlikely the company will consider this move anytime in the coming years given how well capitalized they are today.

This is a pretty small bet for me partly because I view it as a component of a larger basket of financials. I’m encouraged by the value available in the sector and want to spread some of that across various stocks. For more conservative investors, this would be an ideal stock to own and tuck away into the back of your portfolio, only to check on it once a year to make sure the loan book isn’t deteriorating.